From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Jan 18 14:16:01 2005 for ; Tue, 18 Jan 2005 14:16:00 -0500 (EST) for ; Tue, 18 Jan 2005 14:16:00 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 18 Jan 2005 14:15:30 -0500 19:15:30 -0000 j0IJFT7W020362 for ; Tue, 18 Jan 2005 14:15:29 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 18 Jan 2005 14:15:29 -0500 (EST) Date: Tue, 18 Jan 2005 14:15:29 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Welcome to the CSE 463/563 Listserv! To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ WELCOME TO THE CSE 463/563 E-MAIL LISTSERV FOR SPRING 2005! ------------------------------------------------------------------------ 0. The homepage for Prof. Rapaport's course, CSE 463/563, Introduction to Knowledge Representation and Reasoning, Spring 2005, is: http://www.cse.buffalo.edu/~rapaport/563S05.html >From there, you can access the syllabus and the directory of documents. ------------------------------------------------------------------------ 1. If you are reading this email, then you are on the CSE 463/563 Listserv. ------------------------------------------------------------------------ 2. If you are NOT enrolled in the course and therefore do NOT want to be on this Listserv, please either remove your address from it by visiting: http://listserv.buffalo.edu/cgi-bin/wa?SUBED1=cse563-sp05-list&A=1 or by visiting: http://www.cse.buffalo.edu/~rapaport/563S05/listserv.html and following the links for joining or quitting the list, or by sending email to me: rapaport@cse.buffalo.edu ------------------------------------------------------------------------ 3. For information on how to use this Listserv, please visit: http://www.cse.buffalo.edu/~rapaport/563S05/listserv.html ------------------------------------------------------------------------ 4. I will use this mailing list to post important updates, additions, corrections, etc., that I may not have time for in lecture. The TA will use this mailing list for similar purposes. And YOU can use this mailing list to discuss topics from lecture, to ask questions, to get further information, ... If you prefer to contact the TA or me in private (i.e., if you don't want everyone else in class to know what you're asking), then please send us private email (our email addresses are on the syllabus). However, I reserve the right to re-post such email---without your name or any other identifying information---if I think the question and its answer are of general interest. If you do not want me to do this, please let me know in your email message. ------------------------------------------------------------------------ 5. The 5 Commandments of Computer Security: ---------------------------------------- I. Thou shalt maintain the security of thy computer accounts and thy written work. (In plain English: Practice "safe" computing, and don't let anyone copy your work!) II. Thou shalt not share passwords with anyone, nor write thy password down where it may be seen by others. (In plain English: Make sure no one except you knows your password or can find it out!) III. Thou shalt not change permissions to allow others to read thy course directories and files. (In plain English: Don't let anyone else use your account!) IV. Thou shalt not walk away from a workstation without logging out. (In plain English: Always log out!) V. In groups that collaborate inappropriately, it may be impossible to determine who has offered work to others in the group, who has received work, and who may have inadvertantly made their work available to the others by failure to maintain adequate personal security. In such cases, thou shalt all be held equally liable. (In plain English: If your friend copies your work, BOTH of you will be held equally responsible! See Commandment I, above.) ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Jan 19 14:13:13 2005 for ; Wed, 19 Jan 2005 14:13:12 -0500 (EST) for ; Wed, 19 Jan 2005 14:13:12 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 19 Jan 2005 14:12:59 -0500 19:12:59 -0000 j0JJCw7W027914 for ; Wed, 19 Jan 2005 14:12:58 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 19 Jan 2005 14:12:58 -0500 (EST) Date: Wed, 19 Jan 2005 14:12:58 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Commonsense Reasoning To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 3173 ------------------------------------------------------------------------ Subject: CSE 4/563: Commonsense Reasoning ------------------------------------------------------------------------ Just as KRR is a branch of AI, so there are branches of KRR. One of them is called "commonsense reasoning". (Arguably, it was the original branch of KRR: John McCarthy's early paper on KRR was titled "Programs with Common Sense".) The following email notice just came out, and I thought you might be interested in reading the preamble describing this branch of KRR. | CALL FOR PAPERS | | CommonSense 2005 | The 7th International Symposium on Logical | Formalizations of Commonsense Reasoning | | http://www.iccl.tu-dresden.de/commonsense05/ | | May 22-24, 2005 | Corfu, Greece | | | | One of the major long-term goals of artificial intelligence is to endow | computers with commonsense reasoning capabilities. Although we | know how to design and build systems that excel at certain bounded | or mechanical tasks which humans find difficult, such as playing | chess, we have little idea how to construct computer systems that do | well at commonsense tasks which are easy for humans. | Formalizing commonsense reasoning using logic-based approaches | will be the focus of the symposium. Topics of interest include, but | are not limited to: | | * change, action, and causality | * self-aware systems | * axiomatizations of benchmark commonsense problems | * ontologies, including space, time, shape, and matter, and | ontologies of networks and structures | * levels of granularity of ontology and reasoning | * large commonsense knowledge bases | * exploration of new commonsense domains in a preformal way | (e.g. new microworlds or benchmark problems) | * nonmonotonic reasoning | * formal models of probabilistic reasoning | * formal theories of context | * mental attitudes including knowledge, belief, intention, and | planning | * belief revision, update, and merging | * cognitive robotics | * reasoning about multi-agent systems and social interactions | among agents | * applications of formal representations to applications, such as | natural language processing | * other mathematical tools for capturing common-sense | reasoning | | The symposium aims to bring together researchers who have | studied the formalization of commonsense reasoning. The focus of | the symposium is on representation rather than on algorithms, and | on formal rather than informal methods. Papers should be rigorous | and concrete. Technical papers offering new results in the area are | especially welcome; object level theories are preferred. We | especially encourage papers on either of the two themes of this | symposium, Self-awareness and the Surprise Birthday Present | Problem (see below). Survey papers, papers studying the | relationship between different approaches, and papers on | methodological issues such as theory evaluation, are also | encouraged. | | | | Symposium Themes | ================ | There will be two themes at this year's Common Sense Symposium. | In addition to the topics listed above, we encourage papers on these | two themes, and plan to organize one or more panels on these topics. | | The first theme is Self-awareness: the notion of the computer having | a sense of self, being conscious of its own reasoning power, and | being able to explore itself in relation to other agents. We are | especially interested in formal theories that represent self- | awareness, and/or allow a system to reason about its own | awareness. | | The second theme is the Surprise Birthday Present Problem, | one of the challenge problems on the Common Sense Problem | Page (www-formal.stanford.edu/leora/commonsense/). We | encourage submission of papers that present solutions to this | problem and to its listed variants. Sample solutions to other | challenge problems can be found on the Common Sense Problem | Page. | | | | Submission Information | ====================== | Persons wishing to make presentations at the workshop should | submit papers of up to 6000 words, excluding the bibliography. | Electronic submissions in pdf, are preferred; otherwise 6 hard | copies of the paper are acceptable. All submissions should be sent | to one of the Symposium Chairs. | | | | Publication | =========== | The proceedings of the symposium will be published as a Dresden | University Technical Report (ISSN 1430-211X). Moreover, the | program chairs will invite selected authors to submit extended | versions of their papers for a special issue of the Journal of Logic | and Computation. | | | | Multiple Submissions Allowed | ============================ | Papers may be submitted to Commonsense-2005 even if they have | been submitted to other conferences or symposia (such as IJCAI- | 2005). If a paper is accepted at an archival conference such as | IJCAI and is also selected by Commonsense-2005 for publication at | the special journal issue, then this paper must be substantially | revised and/or extended. Previously published papers are not | acceptable for Commonsense-2005. | | | | Participation | ============= | Persons wishing to attend the symposium should submit a 1-2 page | research summary including a list of relevant publications. This is not | required for the authors of submitted papers. Moreover PhD | students need only to send the tentative title and abstract of their | dissertation. All requests for attendance should be sent to one of the | Symposium Chairs. | | | | Important Dates | =============== | Paper Submission Deadline: February 8, 2005 | Paper Notification: March 14, 2005 | Camera Ready Papers Due: March 28, 2005 | Symposium: May 22 - 24, 2005 | | | | Invited Speakers | ================ | Peter Gardenfors, University of Lund | Pat Hayes, University of West Florida | John McCarthy, Stanford University | Leora Morgenstern, IBM Watson Research | | | | Symposium Committee | =================== | Co-Chairs | --------- | Sheila McIlraith | Dept of Computer Science | University of Toronto | 6 King's College Road | Toronto, ON, Canada M4K 2W1 | sheila@cs.toronto.edu | | Pavlos Peppas | Dept of Business Administration | University of Patras | Patras 256 00, Greece | ppeppas@otenet.gr | | Michael Thielscher | Dept of Computer Science | Dresden University of Technology | 01062 Dresden, Germany | mit@inf.tu-dresden.de | | | | Program Committee | =================== | Eyal Amir, Uni. of Illinois at Urbana-Champaign, USA | Mike Anderson, University Of Maryland, USA | Grigoris Antoniou, University of Crete, Greece | Chitta Baral, Arizona State University, USA | Gerd Brewka, University of Leipzig, Germany | Vinay Chaudhri, SRI, USA | Ernie Davis, NYU, USA | Patrick Doherty, University of Linkoping, Sweden | Esra Erdem, Techinal University of Wien, Austria | Nicola Guarino, ISTC-CNR Trento, Italy | Jerry Hobbs, USC/ISI, USA | Antonis Kakas, University of Cyprus, Cyprus | Lefteris Kirousis, University of Patras and CTI, Greece | Jerome Lang, IRIT, France | Vladimir Lifschitz, UT at Austin, USA | Fangzhen Lin, Hong Kong UST | John-Jules Meyer, University of Utrecht, The Netherlands | Leora Morgenstern, IBM Watson Research, USA | Don Perlis, University of Maryland, USA | Fiora Pirri, University of Rome La Sapienza, Italy | Murray Shanahan, Imperial College, UK | Stuart Shapiro, SUNY at Buffalo, USA | Chris Welty, IBM T.J. Watson, USA | Mary-Anne Williams, Uni. of Technology, Sydney, Australia From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Jan 19 22:27:14 2005 for ; Wed, 19 Jan 2005 22:27:13 -0500 (EST) for ; Wed, 19 Jan 2005 22:27:13 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 19 Jan 2005 22:26:51 -0500 03:26:51 -0000 j0K3Qp7W000359 for ; Wed, 19 Jan 2005 22:26:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 19 Jan 2005 22:26:50 -0500 (EST) Date: Wed, 19 Jan 2005 22:26:50 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 563: Syllabus Update To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 3213 ------------------------------------------------------------------------ Subject: CSE 563: Syllabus Update ------------------------------------------------------------------------ I have updated the syllabus to include my office hours, and a reminder about the Shapiro 2003 article on KR. There is also another reading assignment that I've put there that I'll tell you more about next time. For the syllabus, go to: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Jan 20 20:38:10 2005 for ; Thu, 20 Jan 2005 20:38:09 -0500 (EST) for ; Thu, 20 Jan 2005 20:38:09 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 20 Jan 2005 20:38:01 -0500 01:38:01 -0000 Date: Thu, 20 Jan 2005 20:40:21 -0500 Reply-To: Knowledge Representation & Reasoning From: Blake Martin Subject: Book For Course To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: Junk X-UID: 3334 I'd like to save all of you a little time. I went to the College Bookstore and the UB Bookstore today and the book's sold out at both. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Jan 20 21:39:08 2005 for ; Thu, 20 Jan 2005 21:39:07 -0500 (EST) for ; Thu, 20 Jan 2005 21:39:07 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 20 Jan 2005 21:38:53 -0500 02:38:53 -0000 j0L2cq7W007697 for ; Thu, 20 Jan 2005 21:38:52 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 20 Jan 2005 21:38:52 -0500 (EST) Date: Thu, 20 Jan 2005 21:38:52 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: Re: Book For Course To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU I have asked that the bookstore order more copies right away. In the meantime, you might try amazon.com, bn.com, etc., etc. I will see if I can put Ch. 1 on the web. In the meantime, you could read Shapiro 2003 (linked-to on the syllabus) and the Hayes article in the MIT Ency. of Cog. Sci. http://cognet.mit.edu/MITECS/Entry/hayesp.html as reasonable substitutes. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Jan 21 09:45:40 2005 for ; Fri, 21 Jan 2005 09:45:40 -0500 (EST) for ; Fri, 21 Jan 2005 09:45:40 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 21 Jan 2005 09:45:32 -0500 14:45:31 -0000 j0LEjV7W010638 for ; Fri, 21 Jan 2005 09:45:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 21 Jan 2005 09:45:31 -0500 (EST) Date: Fri, 21 Jan 2005 09:45:31 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 463/563: Brachman & Levesque, Pref & Ch. 1, online! To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 463/563: Brachman & Levesque, Pref & Ch. 1, online! ------------------------------------------------------------------------ I have put a PDF copy of the Preface and Ch. 1 of the text on the web. This is not exactly the same version as the published one, since it is from the pre-publication page proofs. Thus, it may differ slightly, contain typos, and is missing the historical section. But it should suffice till either the bookstore gets its act together or you folks can get it from Amazon, Barnes & Noble, or whomever :-) http://www.cse.buffalo.edu/~rapaport/563S05/prefch1.pdf From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Jan 21 15:19:59 2005 for ; Fri, 21 Jan 2005 15:19:59 -0500 (EST) for ; Fri, 21 Jan 2005 15:19:58 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 21 Jan 2005 15:19:45 -0500 20:19:44 -0000 j0LKJi7W013210 for ; Fri, 21 Jan 2005 15:19:44 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 21 Jan 2005 15:19:44 -0500 (EST) Date: Fri, 21 Jan 2005 15:19:44 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 463/563: INTRODUCTORY LECTURE NOTES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 463/563: INTRODUCTORY LECTURE NOTES ------------------------------------------------------------------------ Most of the material I will be lecturing on will be available in various published documents. An exception are my introductory remarks. But I'm hesitant to produce a beautifully formatted and linked webpage (for reasons of time) or to make it public, so I've whipped together an outline of the lectures I've given so far (minus many off-the-cuff side comments) and will be giving over the next lecture or two. It's on the web (but not listed on the Directory of Documents) at: http://www.cse.buffalo.edu/~rapaport/563S05/intro.html It's not self-explanatory, and to appreciate many items, you need to follow appropriate links elsewhere on the course website, but it summarizes what I have been/will be talking about. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Jan 23 22:07:22 2005 for ; Sun, 23 Jan 2005 22:07:22 -0500 (EST) for ; Sun, 23 Jan 2005 22:07:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 23 Jan 2005 22:07:02 -0500 03:07:02 -0000 for ; Sun, 23 Jan 2005 22:07:02 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 23 Jan 2005 22:07:02 -0500 (EST) Date: Sun, 23 Jan 2005 22:07:02 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Private Website To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU The username & password to access the material that needs a username & password on the course website (such as the "Project HALO" article) is: username/login = CSE563 password = Spring2005 Please note that this is case sensitive. If you have trouble accessing anything, please let me know. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 24 14:24:52 2005 for ; Mon, 24 Jan 2005 14:24:51 -0500 (EST) for ; Mon, 24 Jan 2005 14:24:51 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 14:24:48 -0500 19:24:48 -0000 j0OJOlWD006552 for ; Mon, 24 Jan 2005 14:24:47 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 24 Jan 2005 14:24:47 -0500 (EST) Date: Mon, 24 Jan 2005 14:24:47 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Frege, on sense and reference To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: Frege, on sense and reference ------------------------------------------------------------------------ Here is a citation to Frege's paper that I mentioned in lecture today: Frege, Gottlob (1892), "On Sense and Reference," in Peter Geach & Max Black (eds.), Translations from the Philosophical Writings of Gottlob Frege Oxford: Basil Blackwell, 1960. A version (not necessarily the one cited above) is online at: http://mind.ucsd.edu/syllabi/00-01/phil235/a_readings/frege_S&R.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 24 15:17:05 2005 for ; Mon, 24 Jan 2005 15:17:05 -0500 (EST) for ; Mon, 24 Jan 2005 15:17:05 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 15:16:50 -0500 20:16:50 -0000 Date: Mon, 24 Jan 2005 15:19:19 -0500 Reply-To: Knowledge Representation & Reasoning From: Blake Martin Subject: Monads (Philosophy) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU >From www.dictionary.com: monad - An indivisible, impenetrable unit of substance viewed as the basic constituent element of physical reality in the metaphysics of Leibnitz. For (what I'm guessing is) a better description, see http://www.utm.edu/research/iep/l/leib-met.htm#Substance%20as%20Monad My interpretation is that monads are analogous to the human soul regarding many properties (at least according to the Catholic view of the constitution of the human soul; that it is one with no divisible parts; however, the soul is immaterial, whereas monads are partless physical constituents which the world is made of.) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 24 16:43:24 2005 for ; Mon, 24 Jan 2005 16:43:24 -0500 (EST) for ; Mon, 24 Jan 2005 16:43:24 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 16:43:20 -0500 21:43:19 -0000 j0OLhJWD007943 for ; Mon, 24 Jan 2005 16:43:19 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 24 Jan 2005 16:43:19 -0500 (EST) Date: Mon, 24 Jan 2005 16:43:19 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: FINAL EXAM To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: FINAL EXAM ------------------------------------------------------------------------ Our final exam will be: Thursday, May 5, 11:45 a.m. - 2:45 p.m., Knox 4 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 24 16:51:20 2005 for ; Mon, 24 Jan 2005 16:51:19 -0500 (EST) for ; Mon, 24 Jan 2005 16:51:19 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 16:50:00 -0500 21:49:59 -0000 j0OLnxWD007988 for ; Mon, 24 Jan 2005 16:49:59 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 16:49:59 -0500 (EST) Date: Mon, 24 Jan 2005 16:49:59 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: Re: Monads (Philosophy) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: Re: Monads (Philosophy) ------------------------------------------------------------------------ Thanks Blake; another reputable online source is: Kulstad, Mark, & Carlin, Laurence (2002), "Leibniz's Philosophy of Mind", Stanford Encyclopedia of Philosophy http://plato.stanford.edu/entries/leibniz-mind/ From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 24 17:29:34 2005 for ; Mon, 24 Jan 2005 17:29:34 -0500 (EST) for ; Mon, 24 Jan 2005 17:29:34 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 24 Jan 2005 17:28:57 -0500 22:28:57 -0000 X-Originating-IP: 67.23.174.149 Date: Mon, 24 Jan 2005 17:28:57 -0500 Reply-To: Knowledge Representation & Reasoning From: Paula Chesley Subject: Re: CSE 4/563: FINAL EXAM To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU In-Reply-To: <200501242143.j0OLhJRk007942@wasat.cse.buffalo.edu> Quoting "William J. Rapaport" : > ------------------------------------------------------------------------ > Subject: CSE 4/563: FINAL EXAM > ------------------------------------------------------------------------ > > Our final exam will be: > > Thursday, May 5, 11:45 a.m. - 2:45 p.m., Knox 4 > > > That is not only Holocaust Remembrance Day, Feast of the Ascension, Cinco de Mayo, Children's Day (Japan), but also National Day of Prayer (USA). Someone's gotta be objecting to this final on *some* grounds. :) Paula *************** The utilitarian trivialities of their table talk--or rather, of his gloomy monologue--seemed to him positively degrading. He explained at length--fighting her attentive silence, sloshing against the puddles of pauses, abhorring himself--that he had had a long and hard journey; that he slept badly; that he was working on an investigation of the nature of Time, a theme that meant struggling with the octopus of one's own brain. She looked at her wristwatch. --Vladimir Nabokov Ada, or Ardor From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Jan 26 17:16:14 2005 for ; Wed, 26 Jan 2005 17:16:14 -0500 (EST) for ; Wed, 26 Jan 2005 17:16:14 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 26 Jan 2005 17:16:05 -0500 22:06:04 -0000 j0QM64tH004933 for ; Wed, 26 Jan 2005 17:06:04 -0500 (EST) 17:06:04 -0500 (EST) Date: Wed, 26 Jan 2005 17:06:04 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: Extra Credit Problems To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hi all, I have created a webpage for the HW#1-style extra credit problems I gave in lecture today: http://www.cse.buffalo.edu/~ag33/563ec.html The extra credit will go towards the recitation component of your grade. If you choose to do these, they will be due at the beginning of lecture, wednesday FEB 02. Let me know if you have any questions. Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Jan 29 14:54:01 2005 for ; Sat, 29 Jan 2005 14:54:01 -0500 (EST) for ; Sat, 29 Jan 2005 14:54:01 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 29 Jan 2005 14:53:51 -0500 19:53:51 -0000 j0TJrpx3025520 for ; Sat, 29 Jan 2005 14:53:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sat, 29 Jan 2005 14:53:51 -0500 (EST) Date: Sat, 29 Jan 2005 14:53:51 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Buffalo Ontology site To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: Buffalo Ontology site ------------------------------------------------------------------------ This news just in... | Date: Sat, 29 Jan 2005 11:52:41 -0500 | From: "Smith, Barry" | Subject: Buffalo Ontology News | | Dear Friends, | | As you will recall, the National Center for Ontological Research was | founded towards the end of last year with two sites, in Buffalo and in | Stanford. NCOR's beta website is available here: http://ncor.us | | Thanks to the generosity of the College of Arts and Sciences and of the NYS | Center for Bioinformatics and Life Sciences, NCOR's Buffalo site now has | some limited financial and administrative support, which is designed to | promote efforts towards securing long-term funding from outside sources for | high-level ontological research in Buffalo. | | Efforts are currently under way to raise such funds for biomedical ontology | projects involving colleagues in Computer Science, the School of Dental | Medicine, and Roswell Park, as well as with our counterparts in Stanford. A | first NCOR event -- on spatial ontologies in anatomy and geography -- is | also scheduled: | | http://ontology.buffalo.edu/anatomy_GIS You should check out these sites to see what kind of ontological activity is going on here at UB! From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 31 14:05:07 2005 for ; Mon, 31 Jan 2005 14:05:07 -0500 (EST) for ; Mon, 31 Jan 2005 14:05:06 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 31 Jan 2005 14:05:00 -0500 19:04:59 -0000 j0VJ4tx3006331 for ; Mon, 31 Jan 2005 14:04:55 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 31 Jan 2005 14:04:54 -0500 (EST) Date: Mon, 31 Jan 2005 14:04:54 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: CogSci: Shapiro on a new logic for SNePS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: CogSci: Shapiro on a new logic for SNePS ------------------------------------------------------------------------ Although the talk announced below sort of jumps the gun on what I'll be telling you about SNePS in lecture, you might still find this talk to be of interest. It is certainly relevant to KRR! | CENTER FOR COGNITIVE SCIENCE | University at Buffalo, State University of New York | | Wednesday, February 2, 2005 | 2:00 pm - 4:00 pm | 280 Park Hall, North Campus | | | Stuart Shapiro, Ph.D. | Department of Computer Science and Engineering | Center for Cognitive Science | University at Buffalo | | "A LOGIC OF ARBITRARY AND INDEFINITE OBJECTS" | | | A hardcopy of this flyer can be found here: | http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/shapiroannounce.pdf | Please print it out and post it in your department/office. | | Center for Cognitive Science | University at Buffalo, State University of New York | 201 Bell Hall, Buffalo, NY 14260 | Phone: (716) 645-2177 ext. 795, Fax: (716) 645-3825 | Email: ccs-cogsci-contact@buffalo.edu | | Open to the Public. | Refreshments will be served. | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 31 14:35:48 2005 for ; Mon, 31 Jan 2005 14:35:47 -0500 (EST) for ; Mon, 31 Jan 2005 14:35:47 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 31 Jan 2005 14:35:31 -0500 19:35:31 -0000 j0VJZVx3006585 for ; Mon, 31 Jan 2005 14:35:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 31 Jan 2005 14:35:31 -0500 (EST) Date: Mon, 31 Jan 2005 14:35:31 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: How to Run the SNePS Demo To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: How to Run the SNePS Demo ------------------------------------------------------------------------ The two demo files that I ran in class today can be found at: http://www.cse.buffalo.edu/~rapaport/assertions.snepslog http://www.cse.buffalo.edu/~rapaport/queries.snepslog To run them, do the following: 0. save a copy of those files in your local directory 1. Run Lisp (preferably in that directory) 2. Load SNePS 3. Enter SNePSLOG 4. "demo" the files To run lisp: I did it by executing the following at the Unix prompt: > /util/acl62/mlisp But there are better ways; for a complete list, see: http://www.cse.buffalo.edu/sneps/instructions.html To load SNePS: At a Lisp prompt, enter: :ld /projects/snwiz/bin/sneps To enter SNePSLOG (a logic-programming interface to the SNePS KRR system): At the next Lisp prompt, evaluate: (snepslog) To "demo" the files: At the SNePSLOG prompt (a colon ":"), type: : demo "filename" av e.g., : demo "assertions.snepslog" av Notes: "demo" is the function that reads the file and executes the SNePSLOG commands in it; "filename" should be the name you give your local copy of the file; it doesn't have to end in "snepslog", but you must include the quotes in the demo command. If you're running Lisp in a different directory than where you saved your copy of the files, you'll need to give the full pathname. "av" tells SNePSLOG to pause after reading each command but before executing it, and to tell you that it is pausing. AND NOW FOR THE IMPORTANT PART: HOW TO QUIT! When the demo is over, you can continue trying to type in input (though I haven't told you what it should look like--i.e., I haven't given you the syntax; but you can find it in the SNePS manual online at: http://www.cse.buffalo.edu/~jsantore/snepsman/ When you want to quit, do this: 1. Exit SNePSLOG, by typing "lisp" at the colon-prompt: : lisp 2. Now you're back in Lisp; to exit Lisp, type ":ex" at the Lisp prompt: cl-user(4): :ex From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 31 14:50:35 2005 for ; Mon, 31 Jan 2005 14:50:35 -0500 (EST) for ; Mon, 31 Jan 2005 14:50:34 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 31 Jan 2005 14:50:29 -0500 19:50:02 -0000 j0VJo1x3006752 for ; Mon, 31 Jan 2005 14:50:01 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 31 Jan 2005 14:50:01 -0500 (EST) Date: Mon, 31 Jan 2005 14:50:01 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: On Representing Emotion To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: On Representing Emotion ------------------------------------------------------------------------ Earlier in the semester, someone asked about representing emotion. Here are a few links that will give you more information on this topic: 1. As always, a good place to begin is the AI Topics website: AI Topics, "emotion" http://www.aaai.org/AITopics/html/emotion.html 2. The two researchers whose names I mentioned in class are Aaron Sloman and Rosalind Picard (both call the field the study of "affect", which is roughly a synonym for "emotion"): Aaron Sloman's research on "cognition and affect": http://www.cs.bham.ac.uk/research/cogaff/0-INDEX.html Also check out his "homepage portal" at: http://www.cs.bham.ac.uk/~axs/ Rosalind W. Picard's research on "affective computing" http://affect.media.mit.edu/ Also check out her homepage: http://web.media.mit.edu/~picard/ From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Jan 31 16:41:48 2005 for ; Mon, 31 Jan 2005 16:41:48 -0500 (EST) for ; Mon, 31 Jan 2005 16:41:48 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 31 Jan 2005 16:41:30 -0500 21:41:25 -0000 j0VLfPtH011818 for ; Mon, 31 Jan 2005 16:41:25 -0500 (EST) 16:41:24 -0500 (EST) Date: Mon, 31 Jan 2005 16:41:24 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: Monday Recitation (B1) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU To all those in my Monday recitation (B1). We determined today that everyone could make it to SOME recitation (either Mon. or Wed.) if the Monday recitation was moved from its current slot (Mon 3pm-3:50pm) an hour earlier to (Mon 2pm-2:50pm). If you were not in recitation today and have a conflict with BOTH Mon. 2pm-2:50pm (the new B1) AND Wed. 4pm-4:50pm (B2) please let me know (via email, not a newsgroup post) and we can work something out. I will request a new room for the Monday recitation. This may or may not be ready for our next meeting. DO NOT ASSUME THAT WE ARE MEETING @ 2pm UNTIL I MAKE A FORMAL POSTING TO THE NEWSGROUP or DR. RAPAPORT ANNOUNCES IT IN CLASS. Thanks, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Feb 1 11:59:23 2005 for ; Tue, 1 Feb 2005 11:59:22 -0500 (EST) for ; Tue, 1 Feb 2005 11:59:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 1 Feb 2005 11:52:38 -0500 16:52:37 -0000 j11Gqbx3012500 for ; Tue, 1 Feb 2005 11:52:37 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 1 Feb 2005 11:52:37 -0500 (EST) Date: Tue, 1 Feb 2005 11:52:37 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: B&L Ch 2 on line To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU I have put B&L, Ch. 2, in an uncorrected page-proof version, on the web at: http://www.cse.buffalo.edu/~rapaport/563S05/Private/ch2.pdf username/login = CSE563 password = Spring2005 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Feb 1 13:05:30 2005 for ; Tue, 1 Feb 2005 13:05:29 -0500 (EST) for ; Tue, 1 Feb 2005 13:05:29 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 1 Feb 2005 11:38:32 -0500 16:38:32 -0000 j11GcVx3012287 for ; Tue, 1 Feb 2005 11:38:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 1 Feb 2005 11:38:31 -0500 (EST) Date: Tue, 1 Feb 2005 11:38:31 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Textbook Update To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: Textbook Update ------------------------------------------------------------------------ I just called the bookstore. They said that since several of you had come in to special order the text, they would not order any more in larger quantities. There is currently one copy available there. Our departmental secretary tells me that the College Bookstore on Maple might also have copies, since she sends our book orders there, too. I would advise those of you who do not yet have the text or who are not waiting for special orders to arrive to order the book immediately from an online dealer such as amazon.com or bn.com, whoever has the better price and/or delivery time. I will try to put the page-proof version of Ch. 2 (i.e., a version with errors, but still readable) on the website. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Feb 2 09:33:07 2005 for ; Wed, 2 Feb 2005 09:33:07 -0500 (EST) for ; Wed, 2 Feb 2005 09:33:07 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 2 Feb 2005 09:32:51 -0500 14:32:51 -0000 j12EVix3018258 for ; Wed, 2 Feb 2005 09:31:44 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 2 Feb 2005 09:31:44 -0500 (EST) Date: Wed, 2 Feb 2005 09:31:44 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: SYLLABUS UPDATED To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: SYLLABUS UPDATED ------------------------------------------------------------------------ I have updated the syllabus; go to: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html#dates From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Feb 2 09:37:52 2005 for ; Wed, 2 Feb 2005 09:37:52 -0500 (EST) for ; Wed, 2 Feb 2005 09:37:52 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 2 Feb 2005 09:37:40 -0500 14:36:52 -0000 j12Eapx3018286 for ; Wed, 2 Feb 2005 09:36:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 2 Feb 2005 09:36:51 -0500 (EST) Date: Wed, 2 Feb 2005 09:36:51 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: ROOM CHANGE EFFECTIVE TODAY !!!!! To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: ROOM CHANGE EFFECTIVE TODAY !!!!! ------------------------------------------------------------------------ Because of the poor sight lines and visibility in our classroom, I requested (and was granted!) a room change. EFFECTIVE IMMEDIATELY, CSE 463/563 WILL MEET IN: 228 NSC From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Feb 2 09:41:41 2005 for ; Wed, 2 Feb 2005 09:41:41 -0500 (EST) for ; Wed, 2 Feb 2005 09:41:41 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 2 Feb 2005 09:41:09 -0500 14:41:09 -0000 j12Ef8x3018335 for ; Wed, 2 Feb 2005 09:41:08 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 2 Feb 2005 09:41:08 -0500 (EST) Date: Wed, 2 Feb 2005 09:41:08 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: FURTHER SYLLABUS UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: FURTHER SYLLABUS UPDATE ------------------------------------------------------------------------ I have further updated the syllabus to reflect the new room. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 3 08:52:16 2005 for ; Thu, 3 Feb 2005 08:52:15 -0500 (EST) for ; Thu, 3 Feb 2005 08:52:15 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Feb 2005 08:51:56 -0500 13:51:56 -0000 j13Dptx3023773; Thu, 3 Feb 2005 08:51:55 -0500 (EST) 08:51:55 -0500 (EST) Date: Thu, 3 Feb 2005 08:51:55 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: ANOTHER KRR-RELEVANT COGSCI TALK To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: ANOTHER KRR-RELEVANT COGSCI TALK ------------------------------------------------------------------------ Here is the announcement of another CogSci talk that is KRR-relevant. | CENTER FOR COGNITIVE SCIENCE | University at Buffalo, State University of New York | | Wednesday, February 9, 2005 | 2:00 pm - 4:00 pm | 280 Park Hall, North Campus | | | | Selmer Bringsjord, Ph.D. | Chris McEvoy | Department of Cognitive Science | Artificial "Intelligence and Reasoning Laboratory (RAI) | Rensselaer Polytechnic Institute | | | "BUILDING A VIRTUAL PERSON(E) FROM THE "DARK SIDE" | | We describe our general approach to building what we call | advanced synthetic characters (or *bona fide* virtual persons), | within the paradigm of logic-based AI. This approach, based on | our RASCALS architecture, seeks to use a cognitive architecture | for ``mid-level" cognition, and advanced logical systems for | more advanced reasoning-intensive thought. To focus our | general approach, we provide a glimpse of our attempt to bring | to life one particular advanced synthetic character from the | "dark side" --- the character known simply as E (for, as you | may have guessed, evil). Building E entails, among other | things, that we formulate an underlying logico-mathematical | definition of evil, and that we manage to engineer as well an | appropriate presentation of E. | At the presentation level, we use an approach based in | manipulating facial musculature. | | | A hardcopy of this flyer can be found here: | http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/bringsjordannounce.pdf | Please print it out and post it in your department/office. | | | | Center for Cognitive Science | University at Buffalo, State University of New York | 201 Bell Hall, Buffalo, NY 14260 | Phone: (716) 645-2177 ext. 795, Fax: (716) 645-3825 | Email: ccs-cogsci-contact@buffalo.edu | | Open to the Public. | Refreshments will be served. | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 3 11:31:52 2005 for ; Thu, 3 Feb 2005 11:31:52 -0500 (EST) for ; Thu, 3 Feb 2005 11:31:52 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Feb 2005 11:30:04 -0500 16:30:00 -0000 for ; Thu, 3 Feb 2005 11:29:59 -0500 (EST) -0500 (EST) Date: Thu, 3 Feb 2005 11:29:59 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: NEW TIME AND ROOM FOR MONDAY RECITATION B1 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello all, I was able to obtain a new room/time for recitation B1. We will now meet Monday 2pm-2:50pm in Baldy 125 This change is effective immediately (so our next meeting will be in Baldy 125, Mon @ 2pm). Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 3 14:49:44 2005 for ; Thu, 3 Feb 2005 14:49:43 -0500 (EST) for ; Thu, 3 Feb 2005 14:49:43 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Feb 2005 14:49:30 -0500 19:49:30 -0000 j13JnTx3026266 for ; Thu, 3 Feb 2005 14:49:29 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Feb 2005 14:49:29 -0500 (EST) Date: Thu, 3 Feb 2005 14:49:29 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: SHAPIRO'S LOGIC OF ARBITRARY OBJECTS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: SHAPIRO'S LOGIC OF ARBITRARY OBJECTS ------------------------------------------------------------------------ For more information on Shapiro's logic of arbitrary and indefinite objects, and for a copy of the paper that his CogSci talk was based on (Shapiro 2004), go to the webpage for the SNePS-3 project: http://www.cse.buffalo.edu/sneps/Projects/sneps3.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 3 14:55:09 2005 for ; Thu, 3 Feb 2005 14:55:09 -0500 (EST) for ; Thu, 3 Feb 2005 14:55:09 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Feb 2005 14:54:54 -0500 19:54:54 -0000 j13Jsrx3026326 for ; Thu, 3 Feb 2005 14:54:53 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Feb 2005 14:54:53 -0500 (EST) Date: Thu, 3 Feb 2005 14:54:53 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: UPDATED SYLLABUS; SNeRG To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: UPDATED SYLLABUS; SNeRG ------------------------------------------------------------------------ 1. I have updated the syllabus to include the new room/time for recitation B1. Go to: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html#classmtgs 2. Our KRR research group, "SNeRG" (= SNePS Research Group), meets: Mondays, 3-5 p.m., Bell 242. For a schedule of presentations, go to: http://www.cse.buffalo.edu/sneps/Schedules/spring05.html Feel free to leave early or arrive late (the latter is preferable :-) (Note that this Monday's meeting will not be in Bell 242, but will be in the Center for the Arts, Room 266, instead.) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 3 14:59:25 2005 for ; Thu, 3 Feb 2005 14:59:24 -0500 (EST) for ; Thu, 3 Feb 2005 14:59:24 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Feb 2005 14:59:10 -0500 19:59:10 -0000 j13Jx9x3026354 for ; Thu, 3 Feb 2005 14:59:09 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Feb 2005 14:59:09 -0500 (EST) Date: Thu, 3 Feb 2005 14:59:09 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: COLORLESS GREEN IDEAS SLEEP FURIOUSLY To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: COLORLESS GREEN IDEAS SLEEP FURIOUSLY ------------------------------------------------------------------------ I mentioned Noam Chomsky's sentence "Colorless green ideas sleep furiously" as an example of a syntactically well-formed sentence that was not semantically meaningful. (In fact, it appears to be pairwise anomolous: If something is colorless, it can't be green; ideas can't be green; ideas don't sleep; sleeping is not something that can be done furiously.) However, never underestimate the power of the human mind to construct meaning... http://www.cse.buffalo.edu/~rapaport/675w/colorless.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 4 09:26:39 2005 for ; Fri, 4 Feb 2005 09:26:38 -0500 (EST) for ; Fri, 4 Feb 2005 09:26:38 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 4 Feb 2005 09:26:32 -0500 14:26:32 -0000 j14EQVx3000559 for ; Fri, 4 Feb 2005 09:26:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 4 Feb 2005 09:26:31 -0500 (EST) Date: Fri, 4 Feb 2005 09:26:31 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: MEANING OF "REPRESENT" To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MEANING OF "REPRESENT" ------------------------------------------------------------------------ I thought some of you might find the origin of the word "represent" to be of interest. Here is a link to the Oxford English Dictionary's entry on the word. http://www.cse.buffalo.edu/~rapaport/563S05/represent-oed.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Feb 8 12:59:31 2005 for ; Tue, 8 Feb 2005 12:59:30 -0500 (EST) for ; Tue, 8 Feb 2005 12:59:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 8 Feb 2005 12:59:18 -0500 17:59:17 -0000 j18HxHx3024742 for ; Tue, 8 Feb 2005 12:59:17 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 8 Feb 2005 12:59:17 -0500 (EST) Date: Tue, 8 Feb 2005 12:59:17 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: CogSci: Representing Persons To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: CogSci: Representing Persons ------------------------------------------------------------------------ | CENTER FOR COGNITIVE SCIENCE | University at Buffalo, State University of New York | | Wednesday, February 9, 2005 | 2:00 pm - 4:00 pm | 280 Park Hall, North Campus | | Selmer Bringsjord, Ph.D. | Chris McEvoy | Department of Cognitive Science | Artificial "Intelligence and Reasoning Laboratory (RAI) | Rensselaer Polytechnic Institute | | "BUILDING A VIRTUAL PERSON(E) FROM THE "DARK SIDE" | | We describe our general approach to building what we call | advanced synthetic characters (or *bona fide* virtual persons), | within the paradigm of logic-based AI. This approach, based on | our RASCALS architecture, seeks to use a cognitive architecture | for ``mid-level" cognition, and advanced logical systems for | more advanced reasoning-intensive thought. To focus our | general approach, we provide a glimpse of our attempt to bring | to life one particular advanced synthetic character from the | "dark side" --- the character known simply as E (for, as you | may have guessed, evil). Building E entails, among other | things, that we formulate an underlying logico-mathematical | definition of evil, and that we manage to engineer as well an | appropriate presentation of E. | At the presentation level, we use an approach based in | manipulating facial musculature. | | | A hardcopy of this flyer can be found here: | http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/bringsjordannounce.pdf | Please print it out and post it in your department/office. | | | | Center for Cognitive Science | University at Buffalo, State University of New York | 201 Bell Hall, Buffalo, NY 14260 | Phone: (716) 645-2177 ext. 795, Fax: (716) 645-3825 | Email: ccs-cogsci-contact@buffalo.edu | | Open to the Public. | Refreshments will be served. | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 10 17:11:28 2005 for ; Thu, 10 Feb 2005 17:11:27 -0500 (EST) for ; Thu, 10 Feb 2005 17:11:27 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 10 Feb 2005 17:11:12 -0500 22:11:12 -0000 j1AMBCx3010918 for ; Thu, 10 Feb 2005 17:11:12 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 10 Feb 2005 17:11:12 -0500 (EST) Date: Thu, 10 Feb 2005 17:11:12 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: HW 3 QUERY To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 3 QUERY ------------------------------------------------------------------------ A student writes: | | On HW3 question 2, you are asking the truth table for | the proposition: | ((~P ^ (P v Q)) ^ (~Q v R)) | | Is the "R" at the end intentionally and "R" or | is it a typo and should it be "P" instead? It is intentionally an "R", not a typo. You will need 8 rows in your truth table, one for each possible combination of truth values for 3 atomic wffs (2^3=8). There will be a follow-up question or two on the next HW. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 11 14:29:02 2005 for ; Fri, 11 Feb 2005 14:29:02 -0500 (EST) for ; Fri, 11 Feb 2005 14:29:02 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Feb 2005 14:28:51 -0500 19:28:51 -0000 j1BJSpx3016458 for ; Fri, 11 Feb 2005 14:28:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Feb 2005 14:28:51 -0500 (EST) Date: Fri, 11 Feb 2005 14:28:51 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: HW 3 Answers To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 3 ANSWERS AND GRADING SCHEME ------------------------------------------------------------------------ 1. Syntax "English-translation" Semantics ------ ------------------------------- Notation: [[x]] for: the meaning of x [[Apples]] I eat apples [[Book]] There exists a book that lists all other books. [[Delicious]] Every apple that I have ever eaten has been delicious. [[Eat]] I eat apples and pears [[John]] It is possible that John will be tall when he grows up. [[Mary]] Mary believes that John's father is tall. [[Pears]] I eat pears [[Read1]] You have read the book that lists all other books. [[Read2]] You have read any (i.e., some) book. [[Read3]] You have read the book that lists books that do not exist. [[Read4]] You have read every book. [[Tall]] John's father is tall. [[Walk]] I usually like to take a walk. a) Eat b) (Apples ^ Pears) c) (Eat > Walk) NB: ">" for material conditional d) Delicious e) (Apples > Pears) f) Book g) Depends on whether you interpret "unless" as inclusive or exclusive disjunction! Here, I'll represent it inclusively, for convenience. (Note: "-" for negation.): (- Read1 > (- Read2 v Read3)) ^ (Read3 > Read4) h) Tall i) Mary j) John Grading: Syn/Sem: For each atomic proposition (there are 13), 0 = missing 1 = incorrect 2 = partial credit 3 = correct Total = 39 points For each proposition to be represented, 0 = missing 1 = incorrect 2 = partial credit 3 = correct Total = 30 points Subtotal = 69 points ------------------------------------------------------------------------ 2. a) P Q -P (PvQ) (-P^(PvQ)) - - -- ----- ---------- T T F T F T F F T F F T T T T F F T F F (note that (-P^(PvQ)) entails Q) Grading: input columns: 0,1,2,3 intermediate columns: 0,1,2,3 output column: 0,1,2,3 Total = 9 points. b) P Q R -P -Q (PvQ) (-QvR) (-P^(PvQ)) ((-P^(PvQ))^(-QvR)) - - - -- -- ----- ------ ---------- ------------------- T T T F F T T F F T T F F F T F F F T F T F T T T F F T F F F T T T F F F T T T F T T T T F T F T F T F T F F F T T T F T F F F F F T T F T F F (note that last column entails R) Grading: input: 0,1,2,3 intermediate: 0,1,2,3 output: 0,1,2,3 Total = 9 points Subtotal = 18 points. ------------------------------------------------------------------------ 3. Syntax "English translation" semantics ------ ------------------------------- [[Find]] I will be able to find you. [[Forget]] I forget what you look like. [[Follow]] I follow you.(#) [[Hear]] I can hear you. [[ICome]] I come to Mexico. [[Know]] I know what you look like.(+) [[YouFar]] You will be/are far away.(*) [[YouGo]] You go to Mexico. Notes: (#) Arguably, "I follow you" is intended to mean the same as "I come to Mexico". (+) Arguably, "I forget what you look like" is intended to be the negation of "I know what you look like" (*) It is not unreasonable to represent "You will be far away" differently from "You are far away". But doing so will make it harder, later, to show that this argument is valid. a) (YouGo > YouFar) b) (YouFar > -Hear) (note 1) c) (-Hear > -Know) (note 2) d) ((ICome ^ -Know) > -Find) e) ((YouGo ^ ICome) > -Find) Notes: (1) Maybe "when" should not be translated by ">"? But then the logic goes awry. (2) Maybe this should be: (-Hear > Forget), but then the logic goes awry. Grading: Syn/Sem: For each atomic proposition, 0,1,2,3 (total = 24 points) Rep'n: For each molecular proposition, 0,1,2,3 (total = 15 points) Subtotal = 39 points ======================================================================== Grand total = 126 points 463 both 563 A 120-126 A- 113-119 B+ 105-112 B 99-105 B- 92-98 C+ 85-91 C 71-84 43-84 C- 57-70 D+ 43-56 D 21-42 F 0-21 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 11 14:36:42 2005 for ; Fri, 11 Feb 2005 14:36:42 -0500 (EST) for ; Fri, 11 Feb 2005 14:36:41 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Feb 2005 14:36:36 -0500 19:36:35 -0000 j1BJaZx3016552 for ; Fri, 11 Feb 2005 14:36:35 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Feb 2005 14:36:35 -0500 (EST) Date: Fri, 11 Feb 2005 14:36:35 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: ANOTHER PROOF To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: ANOTHER PROOF ------------------------------------------------------------------------ Rather than wait till Monday to give you another example of a natural-deduction propositional-logic proof, I'll give it to you here. In this one, replacing propositional constant wffs "P" and "Q" by metavariables "\alpha" and "\beta" will turn the proof into a derivation of the commutative-law rule of inference that I told you was derivable. (The proof then would be a "subroutine" or "procedure" with parameters instead of constants; any time you needed the commutative law for ^ in a proof, you could then call that "procedure", passing in specific wffs as parameters, and justifying the conclusion by the derived rule of inference.) Show (P^Q) |- (Q^P) 1. (P^Q) : assumption {Show (Q^P): Strategy: show Q, show P, then invoke ^Intro. To show Q, use ^Elim on the assumption. to show P, use ^Elim on the assumption} 2. P : 1, ^Elim 3. Q : 1, ^Elim 4. (Q^P) : 3,2, ^Intro From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 14 17:19:37 2005 for ; Mon, 14 Feb 2005 17:19:36 -0500 (EST) for ; Mon, 14 Feb 2005 17:19:36 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 14 Feb 2005 17:19:24 -0500 22:19:24 -0000 j1EMJNtH029335 for ; Mon, 14 Feb 2005 17:19:23 -0500 (EST) 17:19:23 -0500 (EST) Date: Mon, 14 Feb 2005 17:19:23 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: Homework #4 / V Intro and Elim rules To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello all, I checked with Dr. Rapaport and the inference rules given in HW #4 DO treat P and Q as metavariables (i.e., they can stand in for BOTH atomic and molecular propositions). This is consistent with how we treat them in ^ Introduction and ^ Elimination. So (to the B1 recitation) both natural deduction proofs for the second problem in recitation were correct as written (no negation intro/elim rules needed since -Q can stand in for a metavariable). Also, in HW #4 question #5 the instructions say: "determine which of the above wffs are..." here, "above" refers to only those wffs in Problem #4. Please see me (or send me email) if you have any further questions, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 14 19:08:45 2005 for ; Mon, 14 Feb 2005 19:08:45 -0500 (EST) for ; Mon, 14 Feb 2005 19:08:45 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 14 Feb 2005 19:08:29 -0500 00:08:16 -0000 j1F08Gx3029395 for ; Mon, 14 Feb 2005 19:08:16 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 14 Feb 2005 19:08:16 -0500 (EST) Date: Mon, 14 Feb 2005 19:08:16 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: Re: Homework #4 / V Intro and Elim rules To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: Re: Homework #4 / V Intro and Elim rules ------------------------------------------------------------------------ A follow-up to Albert's posting. I apologize for the confusing way I specified vIntro and vElim on HW4; blame it on cut and paste. As Albert noted, "P" and "Q" in those rules should have been \alpha and \beta, respectively. (And I'm glad you folks noticed this!) I will try to fix the online HW sheet to reflect this, but here's what it should have been: vIntro: From \alpha From \beta ---------------------- ---------------------- Infer (\alpha v \beta) Infer (\alpha v \beta) vElim: From (\alpha v \beta) From (\alpha v \beta) and -\alpha and -\beta --------------------- --------------------- Infer \beta Infer \alpha And, yes, in #5, "above" refers only to 4a--4f. Sorry for the confusion. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 14 20:44:05 2005 for ; Mon, 14 Feb 2005 20:44:05 -0500 (EST) for ; Mon, 14 Feb 2005 20:44:04 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 14 Feb 2005 20:43:54 -0500 01:43:54 -0000 j1F1hrx3000317 for ; Mon, 14 Feb 2005 20:43:53 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 14 Feb 2005 20:43:53 -0500 (EST) Date: Mon, 14 Feb 2005 20:43:53 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: MODAL LOGIC To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MODAL LOGIC ------------------------------------------------------------------------ I mentioned modal logic in lecture in connection with representing "can't". Here's the idea: "It can't be the case that P" (where P is some proposition) roughly means: "It is impossible that P". In modal logic, there are two "operators" that apply to wffs to form new wffs as follows (forgive me: they're hard to write in plain text): <>P ("Diamond P") []P ("Box P") The diamond represents the phrase "it is possible that", and the box represents the phrase "it is necessary that". So, "it can't be the case that P" could be represented as: -<>P (which, by the way, is logically equivalent to: []-P i.e., it is necessary that it is not the case that P.) The semantics involves the notion of a "possible world"; roughly: <>P is true in the actual world iff P is true in some possible world that is "accessible" from the actual world. []P is true in the actual world iff P is true in every possible world that is "accessible" from the actual world. For more information, see the link on the updated FOL webpage: http://www.cse.buffalo.edu/~rapaport/563S05/fol.html#modal or go directly to: http://www.cse.buffalo.edu/~rapaport/663/F03/modallogic.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Feb 15 10:41:34 2005 for ; Tue, 15 Feb 2005 10:41:33 -0500 (EST) for ; Tue, 15 Feb 2005 10:41:33 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 15 Feb 2005 10:41:24 -0500 15:41:24 -0000 j1FFfKx3003632 for ; Tue, 15 Feb 2005 10:41:20 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 15 Feb 2005 10:41:20 -0500 (EST) Date: Tue, 15 Feb 2005 10:41:20 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: HW 4 UPDATED To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 4 UPDATED ------------------------------------------------------------------------ I have updated the online version of HW 4 to include printable metavariables and to make problem 5 more explicit. There is no change in what you have to do; I have merely incorporated the previous clarificatory email messages into the online version of the HW. http://www.cse.buffalo.edu/~rapaport/563S05/hw04.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 18 11:10:35 2005 for ; Fri, 18 Feb 2005 11:10:35 -0500 (EST) for ; Fri, 18 Feb 2005 11:10:34 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 18 Feb 2005 11:10:25 -0500 16:10:25 -0000 j1IGAOjc012164; Fri, 18 Feb 2005 11:10:24 -0500 (EST) User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.0.1) Gecko/20020920 Netscape/7.0 Date: Fri, 18 Feb 2005 11:10:24 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Organization: SUNY Buffalo Computer Science & Engineering Subject: CSE 4/563: Historia Matematica Mailing List Archive: Re: [HM] Modus ponens & tollens To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 3777 ------------------------------------------------------------------------ CSE 4/563: Modus ponens & tollens ------------------------------------------------------------------------ I found a website that translates "Modus Ponens" (and "Modus Tollens", about which you will learn in today's lecture :-) into English: -- ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 20 17:34:59 2005 for ; Sun, 20 Feb 2005 17:34:59 -0500 (EST) for ; Sun, 20 Feb 2005 17:34:58 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Feb 2005 17:34:48 -0500 22:34:48 -0000 j1KMYlx3006562 for ; Sun, 20 Feb 2005 17:34:47 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 20 Feb 2005 17:34:47 -0500 (EST) Date: Sun, 20 Feb 2005 17:34:47 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: Computational Linguistics talk on KRR -- MONDAY! To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A candidate for the computational linguistics position in the LIN dept. will be speaking on Monday. His topic will be how mondadic second-order logic is relevant to computational linguistics. Since mondadic second-order logic can be considered as a KRR system, you might find the presentation interesting. (A logic is "monadic" if the only predicates are 1-place predicates (i.e., properties, not relations). A logic is "second-order" if it allows those predicates to be variables that can be quantified over. FOL only allows *terms* to be variables that can be quantified over. So, whereas in FOL, you can say things like "For all x, if x is a dog, then x is an animal", in SOL, you can also say things like "For all P and Q, if Fido has property P, then Fido has property Q".) Here is the abstract of the talk: ------------------------------------------------------------------------ Second-Order Logic, Finite-State Transducers, and Computational Linguistics Nathan Vaillette University of Tuebingen Monday, Feb. 21, 3:30-4:30 p.m., Baldy 684 Finite-state techniques enjoy a wide popularity in many different domains of Computational Linguistics, such as grapheme-to-phoneme conversion, morphological analysis, partial parsing, and information extraction. Not only do these techniques offer an extremely efficient implementation for many such problems, but furthermore, their flexibility makes them easy to combine and modify, allowing the developer to build solutions modularly and create new applications on the basis of old ones. However, these advantages are often counterbalanced by the relative difficulty of use of finite-state techniques, especially for linguists who are not primarily programmers. In particular, it is often difficult to see the conceptual connection between a linguistic notion and the finite-state machine that implements it. Various formalisms exist to help overcome this difficult. The replacement operator described in (Kaplan and Kay 1994) behaves like the familiar rewrite rules of generative phonology and can be automatically compiled into efficient finite-state machines. The two-level morphological rules of Koskenniemi (1983) constitute an alternative specification language for bridging the gap between linguists' expert knowledge and the computer's efficient finite-state implementation. Nonetheless, these formalisms suffer from serious drawbacks. They each presuppose a certain approach to linguistic problems and are not easily extended. Furthermore, the algorithms compiling the formalisms to finite-state machines are complex and have proven bug-prone. In this presentation, I will discuss an alternative language for simplifying communication between people and finite-state machines: monadic second-order logic (MSOL). This logical language allows humans to express their knowledge to a computer in a way that differs as little as possible from a careful formal presentation of the same information such as they might deliver to other humans. Logic has the further advantage of facilitating formal verifiability: the correctness of a "program" written in it can be mathematically demonstrated, often very straightforwardly. I will demonstrate how MSOL can be used to describe linguistic phenomena and discuss how formulas of the logic can be automatically compiled into efficient finite state machines, building off the pioneering work of Richard Buechi (1960). References: J. Richard Buechi. Weak second order arithmetic and finite automata. Zeitschrift fuer mathematische Logik und Grundlagen der Mathematik, 6:66--92, 1960. Ronald Kaplan and Martin Kay. Regular models of phonological rule systems. Computational Linguistics 20(3):331--378, 1994. Kimmo Koskenniemi. Two-level morphology: a general computational model for word-form recognition and production. Technical report, Department of General Linguistics, University of Helsinki, Finland, 1983. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 20 18:08:30 2005 for ; Sun, 20 Feb 2005 18:08:30 -0500 (EST) for ; Sun, 20 Feb 2005 18:08:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Feb 2005 18:08:23 -0500 23:08:22 -0000 j1KN8Mx3006703 for ; Sun, 20 Feb 2005 18:08:22 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 20 Feb 2005 18:08:22 -0500 (EST) Date: Sun, 20 Feb 2005 18:08:22 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: SYLLABUS UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: SYLLABUS UPDATE ------------------------------------------------------------------------ I have updated the syllbus so that it more accurately reflects what we are doing, when we are doing it, and what you should be reading. Go to: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html#dates From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 20 20:40:34 2005 for ; Sun, 20 Feb 2005 20:40:34 -0500 (EST) for ; Sun, 20 Feb 2005 20:40:34 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Feb 2005 20:40:27 -0500 01:40:26 -0000 j1L1eQx3007156 for ; Sun, 20 Feb 2005 20:40:26 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Feb 2005 20:40:26 -0500 (EST) Date: Sun, 20 Feb 2005 20:40:26 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: Re: CSE 4/563: Computational Linguistics talk on KRR -- MONDAY! To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 3892 Correction: The candidate's topic is monadic second-order logic (not "mondadic" :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 21 09:47:33 2005 for ; Mon, 21 Feb 2005 09:47:32 -0500 (EST) for ; Mon, 21 Feb 2005 09:47:32 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 21 Feb 2005 09:47:18 -0500 14:47:18 -0000 j1LElIx3009530 for ; Mon, 21 Feb 2005 09:47:18 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 21 Feb 2005 09:47:18 -0500 (EST) Date: Mon, 21 Feb 2005 09:47:18 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: REASONING BY SYNTACTIC RULES VS. BY SEMANTIC MENTAL MODELS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: Junk X-UID: 3921 ------------------------------------------------------------------------ Subject: CSE 4/563: REASONING BY SYNTACTIC RULES VS. BY SEMANTIC MENTAL MODELS ------------------------------------------------------------------------ I just came across an excellent article: Johnson-Laird, P.N.; Byrne, Ruth M.J.; & Schaeken, Walter (1992), "Propositional Reasoning by Model", Psychological Review 99(3): 418-439. Besides being a good introduction to Johnson-Laird's theory of mental models, the introductory sections (pp. 418-421) are a superb description of the differences between syntactic reasoning via rules of inference and semantic reasoning via models (mental or otherwise). I have added it to the webpage on Mental Models: http://www.cse.buffalo.edu/~rapaport/mental-models.html and put it online at: http://www.cse.buffalo.edu/~rapaport/563S05/Private/johnsonlaird-byrne-schaeken92.pdf Reminder: To access the "Private" section of our website, use: username login: CSE563 password: Spring2005 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 21 20:44:13 2005 for ; Mon, 21 Feb 2005 20:44:13 -0500 (EST) for ; Mon, 21 Feb 2005 20:44:13 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 21 Feb 2005 20:43:47 -0500 01:43:47 -0000 for ; Mon, 21 Feb 2005 20:43:46 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 21 Feb 2005 20:43:46 -0500 (EST) Date: Mon, 21 Feb 2005 20:43:46 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: HW 4 answers and grading scheme To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW #4 ANSWERS & GRADING SCHEME ------------------------------------------------------------------------ 1. a) Prove: -P ^ (PvQ) |- Q 1. -P ^ (PvQ) : assumption {Show Q: Strategy: use vElim on (PvQ) and -P Therefore, need to show (PvQ), -P Strategy: Use ^Elim on line 1} 2. PvQ : 1, ^Elim 3. -P : 1, ^Elim 4. Q : 2,3, vElim Grading: 0 = missing 24 = seriously illegal 48 = correct up to a point, then errs 72 = syntactically valid proof ------------------------------------------------------------------------ b) Prove: -P ^ (PvQ) ^ (-QvR) |- R 1. -P ^ (PvQ) ^ (-QvR) : assumption { Show R: Strategy: Use vElim on -QvR Therefore, need to show -QvR and --Q (!) To show -QvR, use ^Elim on line 1. To show --Q, use -Intro} 2 . -QvR : 1, ^Elim 3. -P : 1, ^Elim 4. PvQ : 1, ^Elim 5. Q : 3,4, vElim {Show --Q, using -Intro: Enter a subproof with assumption -Q, then derive a contradiction} *6. -Q : temporary assumption *7. Q : send 5 *8. --Q : 6,7,6(!), -Intro {Note: This is a bit tricky! I assumed -Q. I needed to find a wff \beta such that I could derive both \beta and -\beta. I take \beta = Q! I can (trivially) derive Q by sending it in from the main proof on line 5. And I can (very trivially derive -Q, since I have assumed it inside this subproof!. The justification for -Intro normally cites 3 lines of proof: The temporary assumption, \beta, and -\beta. In this very special case, those are lines 6, 7, and 6 again!} 9. --Q : return 8 10. R : 2,9, vElim Note that some students may be tempted at line 5 to jump to the conclusion as follows: 6'. R : 2,5, vElim But the move from: 2. -QvR and 5. Q to 6'. R does not follow the vElim rule of inference, which says: >From (\alpha v \beta) and -\alpha --------------------- Infer \beta because here \alpha = -Q, but then -\alpha = --Q (-\alpha is *not* Q!). Now, you can, if you want *derive* a new rule of inference of the form: >From (-\alpha v \beta) and \alpha ---------------------- Infer \beta (and if a student does that, that's fine; in fact, such a proof is included in the full derivation above, as a special case). Grading: 0 = missing 24 = seriously illegal 48 = correct up to a point, then errs 72 = syntactically valid proof Total = 144 points ======================================================================== 2. col#: 1 2 3 4 5 6 a) P Q -P (PvQ) (-P^(PvQ)) (5>Q) - - -- ----- ---------- ----- T T F T F T T F F T F T F T T T T T F F T F F T Grading: input: 0,4,8,12 intermediate: 0,4,8,12 output: 0,4,8,12 Subtotal = 36 points ------------------------------------------------------------------------ b) 1 2 3 4 5 6 7 8 9 P Q R -P -Q (PvQ) (-QvR) (4^6^7) (8>R) - - - -- -- ----- ------ ------- ----- T T T F F T T F T T T F F F T F F T T F T F T T T F T T F F F T T T F T F T T T F T T T T F T F T F T F F T F F T T T F T F T F F F T T F T F T Grading: input: 0,4,8,12 intermediate: 0,4,8,12 output: 0,4,8,12 Subtotal = 36 points Total = 72 points ======================================================================== 3. 1 2 3 4 5 6 7 8 P Q -P -Q (P^Q) -(P^Q) (-Pv-Q) (6=7) [using "=" for bicond'n'l] - - -- -- ----- ------ ------- ----- T T F F T F F T T F F T F T T T F T T F F T T T F F T T F T T T Grading: input: 0,4,8,12 intermediate: 0,4,8,12 output: 0,4,8,12 Total = 36 points ======================================================================== 4. a) If there is smoke, then there is smoke. b) If there is smoke, then there is fire. c) If, if there is smoke, then there is fire, then, if it's not the case that there is smoke, then it's not the case that there is fire. d) (Either) there is smoke or there is fire or it's not the case that there is fire. e) If there is smoke and there is heat, then there is fire iff (i.e., if and only if) (either) if there is smoke, then there is fire or if there is heat, then there is fire. f) If, if there is smoke, then there is fire, then, if there is smoke and there is heat, then there is fire. Grading: For each sentence: 0 = missing 2 = incorrect 4 = partial credit 6 = correct Total = 36 points. ======================================================================== 5. a) Smoke (Smoke > Smoke) ----- --------------- T T F T tautology ------------------------------------------------------------------------ b) Smoke Fire (Smoke > Fire) ----- ---- -------------- T T T T F F F T T F T T contingent ------------------------------------------------------------------------ c) 1 2 3 4 5 6 7 Smoke Fire -Smoke -Fire (Smoke > Fire) (-Smoke > -Fire) (5>6) ----- ---- ------ ----- -------------- ---------------- ----- T T F F T T T T F F T F T T F T T F T F F F F T T T T T contingent ------------------------------------------------------------------------ d) 1 2 3 4 Smoke Fire -Fire (1v2v3) ----- ---- ----- ------- T T F T T F T T F T F T F F T T tautology ------------------------------------------------------------------------ e) 1 2 3 4 5 6 7 8 9 Smoke Heat Fire (S^H) (S^H)>Fire (S>Fire) (H>Fire) (6v7) (5=8) ----- ---- ---- ----- ---------- -------- -------- ----- ----- T T T T T T T T T T T F T F F F F T T F T F T T T T T T F F F T F T T T F T T F T T T T T F T F F T T F T T F F T F T T T T T F F F F T T T T T tautology ------------------------------------------------------------------------ f) 1 2 3 4 5 6 7 Smoke Fire Heat (S>Fire) (S^H) (S^H)>Fire (4>6) ----- ---- ---- -------- ----- ---------- ----- T T T T T T T T T F T F T T T F T F T F T T F F F F T T F T T T F T T F T F T F T T F F T T F T T F F F T F T T tautology ------------------------------------------------------------------------ Grading: For each sentence: truth table: input: 0,1,2,3 intermediate: 0,1,2,3 output: 0,1,2,3 taut/contra/cont: 0,1,2,3 Subtotal = 12 points/sentence Total = 72 points ======================================================================== Grand total = 288 points 463 both 563 A 273-288 A- 257-272 B+ 241-256 B 225-240 B- 209-224 C+ 193-208 C 161-192 97-192 C- 129-160 D+ 97-128 D 49-96 F 0-48 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Feb 23 18:55:40 2005 for ; Wed, 23 Feb 2005 18:55:40 -0500 (EST) for ; Wed, 23 Feb 2005 18:55:39 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 23 Feb 2005 18:54:32 -0500 23:54:32 -0000 for ; Wed, 23 Feb 2005 18:54:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 23 Feb 2005 18:54:31 -0500 (EST) Date: Wed, 23 Feb 2005 18:54:31 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: FUNCTION SYMBOLS VS. PREDICATES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------- Subject: FUNCTION SYMBOLS VS. PREDICATES ------------------------------------------------------------------------- Since some of you expressed confusion about the difference between function symbols and predicates, I thought I'd summarize what I said in lecture today. If you are NOT confused, you probably shouldn't read this posting, since it may confuse you! But if you ARE confused, maybe it will help. First, don't forget to distinguish between syntax and semantics, in this case, between function symbols and predicates, on the one hand, and functions and properties/relations, on the other hand: SYNTAX SEMANTICS function symbols functions predicates properties and relations Let's concentrate on semantics (or ontology) first: A function is a set of input/output pairs such that the same input always yields the same output (i.e., you'll never have different outputs coming from the same input). A typical function is the doubling function: If the input is an integer, the output is twice the input. Another function is the one that takes as input a person and returns as output the father of that person. A property, mathematically (or "extensionally") speaking, is a set. An n-place relation, mathematically (or "extensionally") speaking, is a set of n-tuples. (Note that, on this view, a property is just a 1-place relation.) A typical example of a property thus understood might be the set of all red things; this would be the property "red". A typical example of a 2-place relation is the set of all ordered pairs of integers such that the first member of the pair is less than the second member; this would be the relation "<". There's another view of properties and relations that you should be aware of: A property, "intensionally" speaking, is some aspect or feature of a thing. So, the property of being red is that feature of a thing that reflects light of a certain wavelength. (Presumably, all the members of the extensional set of red things have this feature.) The An n-place relation, "intensionally" speaking, is some relationship that obtains among the n things. So, the relation of one integer being less than another is the relationship that obtains between the members of the ordered pairs in the extensional "<" relation. The mathematical, or "extensional", version is easier to understand, so I'll stick with that in what follows. I hope it's clear that functions and n-place relations are different things. (But, if you'd like, I can easily confuse you by pointing out that, mathematically, a function is just a 2-place relation, since it's a set of ordered pairs!) Now back to syntax: Function *symbols* are names of functions, and n-place *predicates* are names of n-place relations. So, "father-of" might be the name of one of the functions I mentioned above, and "double" might be the name of the other. We can then write things like this: father-of(George W. Bush) = George H.W. Bush double(2) = 4 Note that what I've just written are wffs of FOL! Each involves the 2-place predicate "=", which forms a wff along with two terms. In the first example, the term on the left-hand side of "=" is a molecular term: father-of(George W. Bush) constructed by taking the atomic constant term "George W. Bush" and forming an expression of the form: f(t). And the term on the right-hand side is an atomic constant term "George H.W. Bush". In the second example, the term on the LHS of "=" is a molecular term: double(2) constructed by taking the atomic constant term "2" and forming an expression of the form: f(t). And the term on the right-hand side is an atomic constant term "4". If I want to express a property of Bush, I could say something like: President(George W. Bush) If the term "George W. Bush" names George W. Bush, and the predicate "President" names the property of being a President of the US (or, mathematically, names the set {George Washington, John Adams, ..., Abraham Lincoln, ..., Bill Clinton, George W. Bush}), then this wff says that Bush has that property. (Just to confuse you, however, some people like to say that predicates name binary functions that take members of the domain as input and outputting a truth value, so "President(George W. Bush)" = true. But I wouldn't want to confuse you :-) The only remaining possible confusion is that the grammatical syntax for functions symbols and predicates is the same: symbol(term1, ..., term-n) But when you formally specify a language for FOL, you must carefully indicate which symbols are to be considered as function symbols and which are predicates. There should never be any ambiguity in a formal language (only in natural languages, like English). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 12:25:59 2005 for ; Thu, 24 Feb 2005 12:25:58 -0500 (EST) for ; Thu, 24 Feb 2005 12:25:58 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 12:25:45 -0500 17:25:45 -0000 j1OHPitH008219 for ; Thu, 24 Feb 2005 12:25:44 -0500 (EST) 12:25:44 -0500 (EST) Date: Thu, 24 Feb 2005 12:25:44 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: On the use of previously derived results in syntactic proofs To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello all, Several of you have asked me whether or not you can use previously derived results as "rules" in a proof. The answer is yes...but I would like you to do this as follows: *If the rule you want to use is a standard rule of inference (the standard connective intro/elim rules from class) then you don't need any further justification (just say something like ^Elim 2 in the comments as usual) *If the rule you want to use is on the same homework (for example if you want to use modus tollens or hypothetical syllogism in your proof of constructive dilemma on HW #5) then, as justification, write "see X" or the name of the rule (if it has one) in the commentary so I know where to look. *If the rule you want to use was proven in class or recitation, then please re-copy it on the side or above your main derivation...DO NOT say something vague like "from class". HOWEVER... You can do all of these proofs with just the basic connective intro/elim rules in hand. Our propositional logic is sound and complete, so if the truth table semantic proof works out, then there MUST be some syntactic derivation using just these inference rules. Also, if you have a derivation in hand that you would like to use for a harder derivation, you could always just substitute for the metavariables as needed to make an appropriate "copy" for your harder derivation (for the programmers out there, this is like expanding a macro). Having said that, I should also say the following: 1c is HARD! It is meant to stretch your logical minds like taffy :)...don't give up on it just because it seems harder than 1a and 1b. Best of luck! Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 13:25:37 2005 for ; Thu, 24 Feb 2005 13:25:37 -0500 (EST) for ; Thu, 24 Feb 2005 13:25:37 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 13:25:18 -0500 18:25:18 -0000 j1OIPIx3001924 for ; Thu, 24 Feb 2005 13:25:18 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 24 Feb 2005 13:25:17 -0500 (EST) Date: Thu, 24 Feb 2005 13:25:17 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: TARSKI ON TRUTH VS. PROOF To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: TARSKI ON TRUTH VS. PROOF ------------------------------------------------------------------------ Alfred Tarski was the logician who developed the first formal definition of "truth". I've put two items on the "Syntax vs. Semantics" webpage relevant to him. The first item is a review of a recent biography of him. The second is a "popular" article he wrote for Scientific American describing the relationships between syntax and semantics, including a sketch of his definition of "truth" and briefly outlining Goedel's incompleteness theorem. Local connection: John Corcoran of the UB PHI department is Tarski's literary executor and editor of a collection of Tarski's papers. Go to: http://www.cse.buffalo.edu/~rapaport/563S05/synsem.html#tarski From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 17:01:45 2005 for ; Thu, 24 Feb 2005 17:01:44 -0500 (EST) for ; Thu, 24 Feb 2005 17:01:44 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 17:01:10 -0500 22:01:09 -0000 for ; Thu, 24 Feb 2005 17:01:09 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 24 Feb 2005 17:01:09 -0500 (EST) Date: Thu, 24 Feb 2005 17:01:09 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: FREEDOM AND BONDAGE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: FREEDOM AND BONDAGE ------------------------------------------------------------------------ Here is a more precise definition of "free" and "bound" occurrences of variables in quantified wffs of FOL. First, there are two slightly different definitions of "scope" floating around in the literature (in what follows, (1) "Q" will be any quantifier, either the universal quantifier (here written as "A") or the existential quantifier (here written as "E"), (2) "v" will be a metavariable ranging over variables, and (3) "Z" and "Y" will be a metavariables ranging over predicates (ordinarily, I'd use "alpha" and "beta", but I can't in this font)): S1: Let Qv[Z] be a wff. Then Z is(df) the scope of (that occurrence of) Q. S2: Let Qv[Z] be a wff. Then Qv[Z] is the scope of (that occurrence of) Q. B&L do not define "scope"! But their definition is consistent with S2 (whereas I had given you S1 in lecture). Following S2, we can then define freedom and bondage of variable occurrences as follows (these definitions are based on Kalish, Montague, & Mar 1980): Def: Let Z be a wff. Then an occurrence of a variable v is bound in Z =df that occurrence stands within an occurrence in Z of a wff of the form QvY. Def: Let Z be a wff. Then an occurrence of a variable v is free in Z =df that occurrence stands within Z but is not bound in Z. I will give you examples in lecture on Friday. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 18:35:27 2005 for ; Thu, 24 Feb 2005 18:35:27 -0500 (EST) for ; Thu, 24 Feb 2005 18:35:27 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 18:34:39 -0500 23:34:39 -0000 j1ONYctH007347 for ; Thu, 24 Feb 2005 18:34:38 -0500 (EST) 18:34:38 -0500 (EST) Date: Thu, 24 Feb 2005 18:34:38 -0500 Reply-To: Knowledge Representation & Reasoning From: Albert Goldfain Subject: subproofs To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Remember that the ONLY reasons to ever enter a subproof are to apply -Intro, -Elim or >Intro. You cannot enter a subproof and make a temp. assumption just to get a consequent of that temp. assumption returned. As I said in recitation, -Intro and -Elim correspond to "Proofs by Contradiction (Reductio Ad Absurdem)"...you assume the opposite of what you are trying to prove, and then show that something "goes wrong" (namely, a pair of contradictory lines \beta, -\beta). >Intro corresponds to "Direct Proofs" (at least this is the term for them in CSE191). You make a temporary assumption \alpha (in a subproof), show that you can legally get to \beta (in the same subproof), and then infer that \alpha > \beta (in the same subproof). Notice that you never return \alpha or \beta outside of the dream (neither would follow logically)...but what you do return is \alpha > \beta. This makes sense semantically...if \alpha is true (and it was as a temp assumption) then \beta is true (because we got to it in the subproof with legal syntactic rules). To extend Dr. Rapaport's "dream" metaphor for subproofs, -Intro and -Elim represent logical "nightmares" (i.e., contradictions) which happen when you enter the dream with the opposite of what you are trying to show. >Intro represents a "good dream" where all of your thoughts link the dream's beginning to it's end. The only dreams you can have are nightmares and good dreams. If this paragraph confuses you then please ignore it :) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 21:30:37 2005 for ; Thu, 24 Feb 2005 21:30:36 -0500 (EST) for ; Thu, 24 Feb 2005 21:30:36 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 21:10:16 -0500 02:10:15 -0000 Date: Thu, 24 Feb 2005 21:13:52 -0500 Reply-To: Knowledge Representation & Reasoning From: Blake Martin Subject: Re: subproofs To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU I understood how the Elim and Intro stuff worked... it seems that none of it applies to problem c, because there isn't a "lone" variable, and a lot of the variables/statements in problem c are symmetrical (in a sense -- I can't think of a better word). The first e-mail helped a lot - I didn't know what the conditions for being able to go into a subproof were. Well, the only way I can see that c can be solved is to assume that one of the consequents is false and show that the other must be true... I can do that for both variables... but that doesn't seem to match any kind of rule that we've encountered so far... it seems more like a "parallel" proof. Well... I guess I'll just write down something... I can see why the conclusion follows, and I can explain why it must hold (in English), but I don't believe I can solve it without some additional form of "proof" ----- Original Message ----- From: "Albert Goldfain" To: Sent: Thursday, February 24, 2005 6:34 PM Subject: subproofs > Remember that the ONLY reasons to ever enter a subproof are to apply > -Intro, -Elim or >Intro. You cannot enter a subproof and make a temp. > assumption just to get a consequent of that temp. assumption returned. > > As I said in recitation, -Intro and -Elim correspond to "Proofs by > Contradiction (Reductio Ad Absurdem)"...you assume the opposite of what > you are trying to prove, and then show that something "goes wrong" > (namely, a pair of contradictory lines \beta, -\beta). > >>Intro corresponds to "Direct Proofs" (at least this is the term for them > in CSE191). You make a temporary assumption \alpha (in a subproof), show > that you can legally get to \beta (in the same subproof), and then infer > that \alpha > \beta (in the same subproof). Notice that you never return > \alpha or \beta outside of the dream (neither would follow > logically)...but what you do return is \alpha > \beta. This makes > sense semantically...if \alpha is true (and it was as a temp assumption) > then \beta is true (because we got to it in the subproof with legal > syntactic > rules). > > To extend Dr. Rapaport's "dream" metaphor for subproofs, -Intro and -Elim > represent logical "nightmares" (i.e., contradictions) which happen when > you enter the dream with the opposite of what you are trying to show. >>Intro represents a "good dream" where all of your thoughts link the > dream's beginning to it's end. The only dreams you can have are > nightmares and good dreams. If this paragraph confuses you then please > ignore it :) > > Albert > From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 22:24:56 2005 for ; Thu, 24 Feb 2005 22:24:54 -0500 (EST) for ; Thu, 24 Feb 2005 22:24:54 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 22:24:17 -0500 03:24:17 -0000 j1P3OGx3003905 for ; Thu, 24 Feb 2005 22:24:16 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 24 Feb 2005 22:24:16 -0500 (EST) Date: Thu, 24 Feb 2005 22:24:16 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: HW 5 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 5 QUESTION ------------------------------------------------------------------------ A student writes: | | I am just having a couple of problems with cse563 HW5 1c First, please note that the night before the HW is due is not a good time to be asking questions. I expect you to begin working on the HW as soon as possible after I assign it, so that you can ask these sorts of questions long *before* the HW is due :-) | I will use the following notation to describe my dilema: a = alpha, b= The word is spelled "dilemma" :-) | beta, g = gamma, l = lambda, d = delta | > = material condition, ^ = and, ~ = not, or = or "conditional" (not "condition" :-) | | QUESTION | if a > b could it's negation be ~a > b or does it have to be ~(a>b) I assume you mean "its" negation :-) [See my howtowrite.html website for the explanation.] The negation of any wff is formed by prefixing the entire wff with the negation sign, so ~(a>b) is the negation of (a>b). | QUESTION | I have tried various ways to do 1c but none of them seem to work here are | the various things that I tried : | | (a or g) ^ (a > b) ^ (g > d) infer (b or d) | 1st way | | 1.)(a or g) ^ (a > b) ^ (g > d) premise | 2.) a or g and elim 1 | 3.) a > b and elim 1 | 4.) g > d and elim 1 | purpose of subroutine to prove a so i can use a and (a > b) to apply >elim | to get b, then use (or intro) to get (b or d) This is reasonable, but not necessarily the best strategy! | *5.) ~a temp. assumpt. | *6.) a or l send 2 I assume you mean "avg" | *7.)g or elim 5,6 | *8.)a > b send 3 | *9.)g > d send 4 | *10.) d > elim 7,9 | *11.)this is where I get stuck because if I use any of the wff from 6 to 8 | as beta then I have to find a ~beta to show a contradiction to prove a or | use the >intro. IF i enter another subroutine then I still have to show a | contradiction to what ever beta I pick and none of the variables or wff | have a negation except alpha. OK--so this is a dead end. | | 2nd way | 1.)(a or g) ^ (a > b) ^ (g > d) premise | 2.) a or g and elim 1 | 3.) a > b and elim 1 | 4.) g > d and elim 1 | purpose of subroutine to return b so I can use or intro on b to say (b or | d) | *5.) a temp. assumpt. | *6.) a > b send 3 | *7.)b >elim 5,6 | * | 8.) stuck here because I can not return b, because last statement in | subroutine must be a negation or >intro rule Right; this is essentially the same problem you ran into above. | also it wouldn't do good to do the following | *8.) g > d send 4 | *9.)a or g send 2 | *9.)a > (g >d) >intro 5,9 I still end up with a molecular formula Right; this won't help. | | 3rd way | 1.)(a or g) ^ (a > b) ^ (g > d) premise | 2.) a or g and elim 1 | 3.) a > b and elim 1 | 4.) g > d and elim 1 | *5.)g temp assumpt. For what purpose? -intro? But you don't need -g | *6.)g > d send 4 | *7.)d >elim 5,6 | *8.) can't do anything with just g and d , stuck here 2 Correct; nothing you can do. | 4th way | 1.)(a or g) ^ (a > b) ^ (g > d) premise | 2.) a or g and elim 1 | 3.) a > b and elim 1 | 4.) g > d and elim 1 | *5.)~g temp. assumpt. For what purpose? -elim? But you don't need g. | *6.)a or g send 2 | *7.)a or elim 5,6 | *8.) stuck here beacuse since I used ~g as my assumption I either ahve to | use ~elim which requires me to show a beta and a ~beta, which I can not do | there is no other ~ other than the one I introduced as alpha and I don't | want to use >intro because because that will return a molecular formula and | I want an atomic one to do either or elim with (a or g) or >elim with (a>b) | OK--you've tried everything except he obvious: Assume -(bvd) and derive a contradiction. Another hint: Try deriving d and -d as the contradiction (I think b and -b will also work). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 22:53:21 2005 for ; Thu, 24 Feb 2005 22:53:21 -0500 (EST) for ; Thu, 24 Feb 2005 22:53:21 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 22:49:39 -0500 03:49:39 -0000 j1P3nZx3004095 for ; Thu, 24 Feb 2005 22:49:35 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 22:49:35 -0500 (EST) Date: Thu, 24 Feb 2005 22:49:35 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: Re: subproofs To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | Date: Thu, 24 Feb 2005 21:13:52 -0500 | From: Blake Martin | | I understood how the Elim and Intro stuff worked... it seems that none of it | applies to problem c, because there isn't a "lone" variable, and a lot of I'm not sure what you mean by a "lone" variable. First, there are no variables; this is a problem in propositional logic, not FOL. But if you mean a wff with a single letter ("P", as opposed to "PvQ", say), then I think you misunderstand the rules of inference. In a rule of inference whose conclusion is, say, "\beta", that "\beta" could be "PvQ". | Well, the only way I can see that c can be solved is to assume that one of | the consequents is false and show that the other must be true... That might be one way, but I doubt it'll work. My suggestion, in my previous posting on this problem, is to assume the entire conclusion is false, and work from there. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Feb 24 22:53:38 2005 for ; Thu, 24 Feb 2005 22:53:38 -0500 (EST) for ; Thu, 24 Feb 2005 22:53:38 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Feb 2005 22:45:16 -0500 03:45:15 -0000 j1P3jFx3004057 for ; Thu, 24 Feb 2005 22:45:15 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 24 Feb 2005 22:45:15 -0500 (EST) Date: Thu, 24 Feb 2005 22:45:15 -0500 Reply-To: Knowledge Representation & Reasoning From: "William J. Rapaport" Subject: CSE 4/563: PROJECT 1 CLARIFICATION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROJECT 1 CLARIFICATION ------------------------------------------------------------------------ Several of you have asked whether you can use the Unix programs "lex" or "yacc" to do part 1 of the project, or whether you can use Prolog to do resolution for you. I've thought about this a bit more, and in both cases the answer is "no". Here's why: Suppose that I've taught you the algorithm for doing long division and asked you to use that algorithm to solve the following division problem: ________ 7)346,788 Now suppose that you solve this problem by getting the answer from a calculator. You've given me the answer, but not by the method I've asked you to practice. And that's the case even if your calculator used the long-division algorithm! Well, using lex, yacc, or Prolog to do the work for you is the same sort of thing. Let me give another analogy: Suppose, again, that I've taught you the long-division algorithm and asked you to implement it in Java. Suppose someone has written a Java method for doing long division, patented it (and kept the algorithm a secret), but distributed it freely on the web as "long-division.java" (or whatever the appropriate name would be; I don't speak Java :-). Now, if you handed in a program that solved the above division problem using "long-division.java" instead of coding your own version of it, that would be just like using a calculator to find the answer. In both cases, it's not the answer I care about; it's whether you understand the algorithm. Here's one more analogy: Now suppose that the code for "long-division.java" is public, and you turn that in instead of writing your own version. That's (almost) OK! I say "almost", because you can't just hand it in, downloaded from the Web or wherever. You need to tell me where you got it, who wrote it, etc. And, to show me that you understand the algorithm, you need to *fully* comment it (every single line of code!). Then it's OK to use it. Now, what about writing a resolution algorithm in Prolog? First, whether or not you know what resolution is (and there's no reason you should, since I haven't taught it to you yet!), here's a fact: Prolog works by using resolution; resolution is built into Prolog. So, using Prolog to do a resolution proof is like using a calculator to do long division. If you really want to use Prolog, feel free: BUT YOU HAVE TO IMPLEMENT RESOLUTION IN PROLOG! (Yes, that means you would have to use resolution to implement resolution!) Here's a (weak) analogy for that: Most programming languages that I know have multiplication built in as a primitive operation. Yet you could (and in some courses are asked to) implement multiplication as either repeated addition or recursively in terms of addition. You would be implementing multiplication in a language that already has multiplication built into it. Here's a stronger analogy: Once in CSE 111 (Great Ideas in Computer Science), I wrote a Pascal program for adding two small integers. That program was then compiled into an assembly-language program by a toy compiler we had running on a Mac. That compiler was, in turn, written in...Pascal! Which, of course, was compiled into a real assembly language for the Mac. Question: Which program was "really" doing the addition? I'm not sure there's a clear answer, but my point is that it's very much like writing a resolution algorithm in a language whose basic underlying method of computing is resolution. It can be done, and, for this project, should be (if you choose to use Prolog, that is). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 25 09:46:53 2005 for ; Fri, 25 Feb 2005 09:46:52 -0500 (EST) for ; Fri, 25 Feb 2005 09:46:52 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 25 Feb 2005 09:46:18 -0500 14:46:18 -0000 j1PEkIx3006251 for ; Fri, 25 Feb 2005 09:46:18 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 25 Feb 2005 09:46:18 -0500 (EST) Date: Fri, 25 Feb 2005 09:46:18 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 5 again To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 5 again ------------------------------------------------------------------------ A student writes: | | The ~(PvQ) approach was the first one I tried... and I can see how it'd be | useful IF we could say ~(PvQ)=(~P^~Q) but I haven't found a way to go from | ~(PvQ) to anything useful. You can only have one temporary assumption, | right? Right, but you can make others. In particular, you could try to prove the equivalence you cite above. Actually, all you need to do is prove each of -P and -Q. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 25 12:43:38 2005 for ; Fri, 25 Feb 2005 12:43:38 -0500 (EST) for ; Fri, 25 Feb 2005 12:43:38 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 25 Feb 2005 12:42:57 -0500 17:42:57 -0000 j1PHgvx3007830 for ; Fri, 25 Feb 2005 12:42:57 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 25 Feb 2005 12:42:57 -0500 (EST) Date: Fri, 25 Feb 2005 12:42:57 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MATERIAL CONDITIONAL To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MATERIAL CONDITIONAL ------------------------------------------------------------------------ For those of you still bothered by the truth table for the material conditional (and why, for that matter, it's called the "material" conditional), here are some answers: First, one reason that "horseshoe" (>) is called the "material" conditional is to contrast it with: subjunctive conditional: if I were to sing, then you would be unhappy counterfactual conditional: if I had sung, then you would have been unhappy You can find out a lot more about the material (and other) conditionals by doing a Google search on "material conditional": http://www.google.com/search?hl=en&q=material+conditional From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 25 12:45:35 2005 for ; Fri, 25 Feb 2005 12:45:35 -0500 (EST) for ; Fri, 25 Feb 2005 12:45:35 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 25 Feb 2005 12:44:59 -0500 17:44:59 -0000 j1PHixx3007853 for ; Fri, 25 Feb 2005 12:44:59 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 25 Feb 2005 12:44:59 -0500 (EST) Date: Fri, 25 Feb 2005 12:44:59 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: AN INVALID ARGUMENT To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: AN INVALID ARGUMENT ------------------------------------------------------------------------ Is this a valid proof in our natural-deduction system for propositional logic? If not, why not? 1. P>Q : assumption *2. P : temporary assumption *3. P>Q : send 1 *4. Q : 3,2,>Elim 5. Q : return 4 The answer will be in the next listserv posting :-) From rapaport@cse.Buffalo.EDU Fri Feb 25 14:13:14 2005 for ; Fri, 25 Feb 2005 14:13:14 -0500 (EST) for ; Fri, 25 Feb 2005 14:13:14 -0500 (EST) Fri, 25 Feb 2005 14:13:13 -0500 (EST) Fri, 25 Feb 2005 14:13:13 -0500 (EST) Date: Fri, 25 Feb 2005 14:13:13 -0500 (EST) From: "William J. Rapaport" To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU, rapaport@cse.Buffalo.EDU Subject: Re: CSE 4/563: AN INVALID ARGUMENT ------------------------------------------------------------------------ Subject: Re: CSE 4/563: AN INVALID ARGUMENT ------------------------------------------------------------------------ I asked: | Is this a valid proof in our natural-deduction system for propositional | logic? If not, why not? | | 1. P>Q : assumption | *2. P : temporary assumption | *3. P>Q : send 1 | *4. Q : 3,2,>Elim | 5. Q : return 4 Here's the answer (don't read it till you've tried answering it!) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Answer: No; it's not valid. You can't return Q to line 5, because >Elim is not one of the legal rules for use of the Rule of Return. See: http://www.cse.buffalo.edu/~rapaport/563S05/sendreturn.html for the correct rule. >From line *4, you could have as the next lines: *5. P>Q : 2,4,>Intro 6. P>Q : return 5 But this gets you nowhere, since you already had P>Q on line 1. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 25 14:43:31 2005 for ; Fri, 25 Feb 2005 14:43:30 -0500 (EST) for ; Fri, 25 Feb 2005 14:43:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 25 Feb 2005 14:42:34 -0500 19:42:34 -0000 j1PJgYx3008419 for ; Fri, 25 Feb 2005 14:42:34 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 25 Feb 2005 14:42:34 -0500 (EST) Date: Fri, 25 Feb 2005 14:42:34 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PROPER SUBPROPOSITION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROPER SUBPROPOSITION ------------------------------------------------------------------------ In lecture, I have several times used the phrase "proper subproposition" (or synonyms of that). Here is a webpage that formally defines it: http://www.cse.buffalo.edu/~rapaport/563S05/propersubwff.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Feb 25 21:05:00 2005 for ; Fri, 25 Feb 2005 21:04:59 -0500 (EST) for ; Fri, 25 Feb 2005 21:04:59 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 25 Feb 2005 21:04:26 -0500 02:04:26 -0000 for ; Fri, 25 Feb 2005 21:04:25 -0500 (EST) Date: Fri, 25 Feb 2005 21:04:24 -0500 Reply-To: stock@BUFFALO.EDU From: Matt Stock Subject: Re: CSE 4/563: PROJECT 1 CLARIFICATION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU In-Reply-To: <200502250345.j1P3jFwC004056@wasat.cse.buffalo.edu> On Feb 24, 2005, at 10:45 PM, William J. Rapaport wrote: > Now suppose that you solve this problem by getting the answer from > a calculator. You've given me the answer, but not by the method > I've asked you to practice. And that's the case even if your > calculator used the long-division algorithm! > > Well, using lex, yacc, or Prolog to do the work for you is the same > sort of thing. If we use lex/yacc/sablecc or other tool to just do the input parsing, would that be fair use of such tools? We clearly need to do the clause manipulation (CNF, skolemization, etc) in our own code, but it seems like these tools will provide the same level of support that lisp provides natively. As for me, I'm likely to use lisp anyway - it's worth relearning it for later projects. :-) -Matt From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Feb 26 10:02:10 2005 for ; Sat, 26 Feb 2005 10:02:09 -0500 (EST) for ; Sat, 26 Feb 2005 10:02:09 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 26 Feb 2005 10:01:46 -0500 15:00:36 -0000 j1QF0Zx3011906; Sat, 26 Feb 2005 10:00:36 -0500 (EST) 10:00:35 -0500 (EST) Date: Sat, 26 Feb 2005 10:00:35 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: CSE 4/563: PROJECT 1 CLARIFICATION Comments: To: stock@BUFFALO.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | Date: Fri, 25 Feb 2005 21:04:24 -0500 | From: Matt Stock | ... | If we use lex/yacc/sablecc or other tool to just do the input parsing, | would that be fair use of such tools? We clearly need to do the clause | manipulation (CNF, skolemization, etc) in our own code, ... Yes; feel free to use a tool for input parsing, but you must implement your own clause-form algorithm. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 27 20:42:47 2005 for ; Sun, 27 Feb 2005 20:42:47 -0500 (EST) for ; Sun, 27 Feb 2005 20:42:47 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Feb 2005 20:41:40 -0500 01:41:40 -0000 for ; Sun, 27 Feb 2005 20:41:40 -0500 (EST) -0500 (EST) Date: Sun, 27 Feb 2005 20:41:39 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Grading for HW #4 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello all, There was an error in the grading chart posted for HW#4 (which you will be receiving back this week in recitation). The total number of points was 360...not 288. Here is the updated grading scheme Grand total = 360 points 463 both 563 A 341-360 A- 321-340 B+ 301-320 B 281-300 B- 261-280 C+ 241-279 C 201-240 121-240 C- 161-200 D+ 121-160 D 61-120 F 0-60 Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 27 21:21:52 2005 for ; Sun, 27 Feb 2005 21:21:51 -0500 (EST) for ; Sun, 27 Feb 2005 21:21:51 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Feb 2005 21:18:02 -0500 02:18:02 -0000 j1S2I1x3017498 for ; Sun, 27 Feb 2005 21:18:01 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 27 Feb 2005 21:18:01 -0500 (EST) Date: Sun, 27 Feb 2005 21:18:01 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: Translating from Propositional Logic to English To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: Translating from Propositional Logic to English ------------------------------------------------------------------------ 1. I've had some questions about how to translate complicated molecular wffs into English. It's actually quite straightforward, even though the results might not sound very idiomatic. (But, after all, very few people except logicians and lawyers talk this way!) Here's an example: Let [[snow]] = It's snowing. and [[sunny]] = It's sunny. and [[buffalo]] = We're in Buffalo. Then: [[ (snow > buffalo) > ((sunny > buffalo) > buffalo)) ]] = if [[ (snow > buffalo) ]] then [[((sunny > buffalo) > buffalo)]] = if, if [[snow]] then [[buffalo]], then, if [[(sunny > buffalo)]] then [[buffalo]] = if, if it's snowing, then we're in Buffalo, then, if, if [[sunny]] then [[buffalo]], then we're in Buffalo = if, if it's snowing, then we're in Buffalo, then, if, if it's sunny, then we're in Buffalo, then we're in Buffalo And that, believe it or not, is grammatically correct English. (Well, it's grammatically correct "logical" English, to be more precise.) The excerpts from Kalish, Montague & Mar 1980 discuss this in detail. (As an aside, I'll certainly agree that it's not "acceptable" English. It would be very difficult to understand a sentence like this without pencil and paper to help analyze it. I'll give you another example of a grammatically correct, but hard to understand, sentence of English at the end of this posting.) 2. As I mentioned in lecture, contrary to what many might think, the phrase "either P or Q" does NOT mean exclusive disjunction (i.e., P or Q, but not both). It is, in fact, ambiguous between the inclusive and exclusive readings. So, when trying to decide how to represent an "either/or" statement in logic, you could use your understanding of the subject matter to decide, but if that doesn't help, you should always err on the side of caution and interpret it *inclusively*. 3. This is not directly related to any of the above, but here are two examples of grammatically correct and meaningful, but hard-to-understand, sentences of English. Can you figure out what they mean? Dogs dogs dog dog dogs. and Buffalo buffalo buffalo buffalo buffalo. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 27 21:26:48 2005 for ; Sun, 27 Feb 2005 21:26:48 -0500 (EST) for ; Sun, 27 Feb 2005 21:26:48 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Feb 2005 21:26:14 -0500 02:26:14 -0000 j1S2QDx3017531 for ; Sun, 27 Feb 2005 21:26:13 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 27 Feb 2005 21:26:13 -0500 (EST) Date: Sun, 27 Feb 2005 21:26:13 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: TRANSLATIONS FROM ENGLISH TO F.O.L. To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------- Subject: TRANSLATIONS FROM ENGLISH TO F.O.L. ------------------------------------------------------------------------- In HW #3, I asked you to represent the following sentences in our language for propositional logic: A good way to practice for the midterm exam's section on FOL translation would be to try to represent these in our language for FOL: (answers in the next posting) 1. I eat apples and pears. 2. I eat apples and I eat pears. 3. If I eat apples and pears, I usually like to take a walk. 4. Every apple that I have ever eaten has been delicious. 5. I eat apples only if I eat pears. 6. There exists a book that lists all other books. 7. If you haven't read the book that lists all other books, then you haven't read any book, unless you've read the book that lists books that do not exist, in which case you've read every book. 8. John's father is tall. 9. Mary believes that John's father is tall. 10. It is possible that John will be tall when he grows up. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 27 22:47:12 2005 for ; Sun, 27 Feb 2005 22:47:10 -0500 (EST) for ; Sun, 27 Feb 2005 22:47:09 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Feb 2005 22:46:23 -0500 03:46:23 -0000 j1S3kNx3017835 for ; Sun, 27 Feb 2005 22:46:23 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 27 Feb 2005 22:46:23 -0500 (EST) Date: Sun, 27 Feb 2005 22:46:23 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563:CSE 4/563: ANSWERS TO HW#3 PROBLEMS IN F.O.L. To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4002 ------------------------------------------------------------------------- Subject: TRANSLATIONS FROM ENGLISH TO F.O.L. ------------------------------------------------------------------------- In HW #3, I asked you to represent the following sentences in our language for propositional logic: A good way to practice for the midterm exam's section on FOL translation would be to try to represent these in our language for FOL: Here are suggested answers, using "pretend-it's-English" semantics: 1. I eat apples and pears. This is actually rather trick to represent in FOL. One possibility is to use a 3-place "Easts" predicate: Eat(i, apples, pears) Or, one could use a 2-place "Eats" predicate (which should probably be called something like "Eats2" to distinguish it from the predicate above) and then introduce set theory, writing: Eat(i, {apples, pears}) But that goes beyond FOL. Another way to go beyond FOL would be to use "mereology" (the logical theory of parts and wholes), which has a "fusion" operator that takes two individuals and produces a new individual that is the whole whose parts are the original two individuals: Eat(i, apples+pears) Or, of course, one could interpret (1) as if it were synonymous with (2): 2. I eat apples and I eat pears. Eat(i, apples) ^ Eat(i,pears). Of course, in both cases, it might be better to say something like: Ax[Apple(x) > East(i,x)] (etc.) But this raises the problem--again, beyond the scope of FOL, though an important topic in KRR ("default reasoning")--that I don't eat all of the apples in the universe! 3. If I eat apples and pears, I usually like to take a walk. Here, let's suppose that "I eat apples and pears" is represented as "Eat(i, apples) ^ Eat(i, pears)". We might represent "I loke to take a walk" as: Like(i, take-a-walk) Not very illuminating. You might be tempted to say something like: Like(i, Take(i,walk)) But what is the syntax and semantics of "Take(x,y)"? In FOL, the only way that could appear as an argument of a predicate would be it were a term, which would require "Take" to be a function symbol, but it's hard to see how "Take(i,walk)" could be interpreted as a noun phrase. I think that the best we can do in FOL is: Like-to-take-a-walk(i) so that the whole sentence would look like this: (Eat(i,apples) ^ Eat(i,pears)) > Like-to-take-a-walk(i) 4. Every apple that I have ever eaten has been delicious. Ax[(Apple(x) ^ Eaten(i,x)) > Delicious(x)] But I urger you to think carefully about the semantics of these predicates if this is intended to be a translation of (4)! 5. I eat apples only if I eat pears. Eat(i,apples) > Eat(i,pears) 6. There exists a book that lists all other books. Ex[Book(x) ^ Ay[(Book(y) ^ -(y=x)) > Lists(x,y)]] 7. If you haven't read the book that lists all other books, then you haven't read any book, unless you've read the book that lists books that do not exist, in which case you've read every book. This is a great challenge, since it requires an interpretation of "the book...", which we discussed in lecture (Russell's interpretion). There is a way to extend FOL to include a "definite description operator" (usually denoted by lower-case Greek iota, but at least one logician, Dana Scott (also a well-known computer scientist), suggests using an inverted sans-serif capital "I" :-). I'll use "the" as the operator. Here's its definition: P((the x).Q(x)) =def Ex[Q(x) ^ Ay[Q(y) > y=x] ^ P(x)] Using this operator, (7) might become (assuming for convenience that "you" means you, and is not an idiomatic way of introducing a universal quantifier ranging over people, and also assuming that "Lists-all-other-books" is a predicate defined as in (6) above :-)): (-Read(you, (the x)[Book(x) ^ Lists-all-other-books(x)]) > (-Ey[Book(y) ^ Read(you,y)] xor Read(you, (the z)[Book(z) ^ Eu[Book(u) ^ -Exists(u) ^ Lists(z,u)]]))) ^ Read(you, (the z)[Book(z) ^ Eu[Book(u) ^ -Exists(u) ^ Lists(z,u)]]) > Av[Book(v) > Read(you,v)] (or so I think!) 8. John's father is tall. Tall(father-of(john)) 9. Mary believes that John's father is tall. This can't be represented in FOL, because if "Believes" is a 2-place predicate taking a believer and a proposition as arguments, then it goes beyond the capabilities of FOL, since propositions can't be arguments of predicates! It requires either "epistemic" logic (a form of modal logic that has a "believes-that" modal operator) or...SNePS :-) (SNePS allows propositions to be "first-class objects"; we'll discuss this later.) 10. It is possible that John will be tall when he grows up. Again, this goes beyond FOL, requiring at least a modal operator for "it is possible that" and perhaps even a temporal operator to represent the future tense: Possible(Future(Tall(john))) might represent "it is possible that, in the future, John is tall". Note that I have not tried to represent "when he grows up". I'll leave that as an exercise for the reader :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 00:14:50 2005 for ; Mon, 28 Feb 2005 00:14:49 -0500 (EST) for ; Mon, 28 Feb 2005 00:14:49 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 00:14:18 -0500 05:14:18 -0000 for ; Mon, 28 Feb 2005 00:14:17 -0500 (EST) -0500 (EST) Date: Mon, 28 Feb 2005 00:14:17 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Quiz Grading Scheme To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Here is the grading scheme for last week's recitation quiz Premises: Knowing that every premise was useful in proving result 0 = didn't take quiz 1 = didn't use all of the premises 2 = used all premises but didn't include them as lines in the proof 3 = listed and used all premises Subproof: Knowing that subproof was needed for >Intro 0 = didn't take quiz 1 = didn't use subproof 2 = used subproof without temp. assumption to begin or without send/return 3 = used a subproof with temp. assumption and proper send/return Syntactic Proof 0 = didn't take quiz 4 = seriously illegal 8 = correct up to a point, then errs 12 = syntactically valid proof Total = 18 points cse463 both cse563 A 18 A- 17 B+ 16 B 15 B- 14 C+ 13 C 11-12 7-12 C- 9-10 D+ 7-8 D 4-6 F 0-3 Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 00:32:27 2005 for ; Mon, 28 Feb 2005 00:32:27 -0500 (EST) for ; Mon, 28 Feb 2005 00:32:27 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 00:25:11 -0500 05:25:10 -0000 for ; Mon, 28 Feb 2005 00:25:10 -0500 (EST) -0500 (EST) Date: Mon, 28 Feb 2005 00:25:09 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: one more thing about thing about the quiz To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU The solution to the quiz (for both recitations) will look an awful lot like hw #5 problem 2 (the mexico argument) OutOf(Cat,Bag) :) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 09:45:29 2005 for ; Mon, 28 Feb 2005 09:45:29 -0500 (EST) for ; Mon, 28 Feb 2005 09:45:29 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 09:44:42 -0500 14:44:41 -0000 j1SEifx3020319 for ; Mon, 28 Feb 2005 09:44:41 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Feb 2005 09:44:41 -0500 (EST) Date: Mon, 28 Feb 2005 09:44:41 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CogSci talk on context To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Although this week's CogSci colloquium is not directly on KRR, you might be interested in it for other reasons ;-) | CENTER FOR COGNITIVE SCIENCE | University at Buffalo, The State University of New York | | | Wednesday, March 2, 2005 | 2:00 pm - 4:00 pm | 280 Park Hall, North Campus | | | William Rapaport, Ph.D. | Department of Computer Science & Engineering | Department of Philosophy | Center for Cognitive Science | | | "In Defense of Contextual Vocabulary Acquisition: | How to Do Things with Words in Context" | | "Context" is notoriously vague, and its uses multifarious. Researchers | in "contextual vocabulary acquisition" differ over the kinds of context | involved in vocabulary learning, and the methods and benefits thereof. | This talk presents a computational theory of contextual vocabulary | acquisition, identifies the relevant notion of context, exhibits the | assumptions behind some classic objections, and defends our theory | against these objections. | | References: | | Beck, Isabel L.; McKeown, Margaret G.; & McCaslin, Ellen S. (1983), | "Vocabulary Development: All Contexts Are Not Created Equal", | Elementary School Journal 83(3): 177-181. | http://ublib.buffalo.edu/cgi-bin/reserve.cgi?B029442831.PDF | | Schatz, Elinore Kress, & Baldwin, R. Scott (1986), | "Context Clues Are Unreliable Predictors of Word Meanings", | Reading Research Quarterly 21(4, Fall): 439-453. | http://ublib.buffalo.edu/cgi-bin/reserve.cgi?B029441932.PDF | | Rapaport, William J. (submitted, 2004), | "In Defense of Contextual Vocabulary Acquisition: | How to Do Things with Words in Context", | submitted to Context-05. | http://www.cse.buffalo.edu/~rapaport/Papers/paris.pdf | | | | A hardcopy of this announcement can be found here: | http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/rapaportannounce.pdf | Pleas print it out and post it in your department and office. | | Center for Cognitive Science | University at Buffalo, The State University of New York | 201 Bell Hall, Buffalo, NY 14260 | Phone (716) 645-2177 ext. 795, Fax: (716) 645-3464 | URL: http://www.cogsci.buffalo.edu | | Open to the public. | Refreshments will be served. | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 14:16:35 2005 for ; Mon, 28 Feb 2005 14:16:34 -0500 (EST) for ; Mon, 28 Feb 2005 14:16:34 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 14:15:31 -0500 19:15:31 -0000 j1SJFUx3023671 for ; Mon, 28 Feb 2005 14:15:30 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Feb 2005 14:15:30 -0500 (EST) Date: Mon, 28 Feb 2005 14:15:30 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: ANSWERS TO HW #5 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ HW #5 ANSWERS & GRADING SCHEME ------------------------------------------------------------------------ 1. Note: Here and below, I use "A" for \alpha, "B" for \beta, "G" for \gamma, and "D" for \delta; and ">" for the material conditional. I omit unnecessary parentheses. a) Proof of Modus Tollens: 1. A>B : assumption 2. -B : assumption {Show -A. Strategy: Use -Intro. Assume A, and derive a contra. Then invoke -Intro.} *3. A : temporary assumption for -Intro *4. A>B : send 1 *5. B : 4,3,>Elim *6. -B : send 2 *7. -A : 3,5,6, -Intro 8. -A : return 7 Note: The above is not, strictly speaking, a proof; rather, it is a proof *schema*, since "A" and "B" are not wffs, but metavariables ranging over wffs. The above schema can be turned into a proof by "binding" specific wffs to the metavariables. Thus, if you are in the middle of a proof, have 2 wffs such as "P>Q" and "-P", and would like to infer "-Q", you could either repeat the above steps with "P" and "Q" replacing (or being bound to) "A" and "B", or else simply infer "-Q", citing MT (as a macro call, or a call to a subroutine; use your favorite metaphor) as the justification, since you have now proved it. Grading: 0 = missing 1 = seriously illegal 2 = correct up to a point, then errs 3 = syntactically valid proof ------------------------------------------------------------------------ b) Proof of Hypothetical Syllogism 1. A>B : assumption 2. B>G : assumption {Show A>G. Strategy: Temporarily assume A, and show G. Then invoke >Intro.} *3. A : temporary assumption for >Intro {Show G. Strategy: Infer G from B>G via >Elim. Therefore, need to show B. Strategy to show B: Infer B from A>B via >Elim. Therefore, need A. But we have A!} *4. A>B : send 1 *5. B : 4,3,>Elim *6. B>G : send 2 *7. G : 6,5,>Elim *8. A>G : 3,7,>Intro 9. A>G : return 8 Grading: 0 = missing 1 = seriously illegal 2 = correct up to a point, then errs 3 = syntactically valid proof ------------------------------------------------------------------------ c) Proof of Constructive Dilemma 1. AvG : assumption 2. A>B : assumption 3. G>D : assumption {Show BvD. Strategy: Note: Can't use vIntro, since then would need to show B (or D), but we can't. Alternative strategy: Try -Elim, by assuming -(BvD) and deriving a contradiction.} * 4. -(BvD) : temporary assumption for -Elim {strategy: infer -B, then use MT to infer -A, then use vElim to infer G, then use >Elim to infer D, then infer -D in the same way that -B was inferred but that contradicts D} {Ah, but how do we infer -B^-D, which follows from 4 by one of DeMorgan's Laws? Well, we *could* do a little truth table on the side, then invoke the completeness of propositional logic to infer that there must be a proof. But that's also like Russell's "theft over honest toil". So, we need to *prove* it! Actually, we only need to prove each of -B and -D; we won't need their conjunction.} {Show -B} ** 5. B : temporary temporary assumption for -Intro ** 6. BvD : 5,vIntro ** 7. -(BvD) : send 4 ** 8. -B : 5,6,7,-Intro * 9. -B : return 8 {Now show -D in the same way} **10. D : temporary temporary assumption for -Intro **11. BvD : 10,vIntro **12. -(BvD) : send 4 **13. -D : 10,11,12,-Intro *14. -D : return 13 {Finally, still in the subproof we entered with temporary assumption -(BvD) on line 4, now derive a contradiction, so that we can eliminate the negation} *15. A>B : send 2 *16. -A : 15,9,MT (alternatively, I could have repeated the entire proof of MT here, instead of using this macro or procedure call) *17. AvG : send 1 *18. G : 17,16,vElim *19. G>D : send 3 *20. D : 19,18,>Elim *21. BvD : 4,20,14,-Elim 22. BvD : return 21 Whew! Grading: 0 = missing 1 = seriously illegal 2 = correct up to a point, then errs 3 = syntactically valid proof Total = 9 points ------------------------------------------------------------------------ 2. Syntax "English translation" semantics ------ ------------------------------- [[Find]] I will be able to find you. [[Follow]] I follow you. [[Hear]] I can hear you. [[ICome]] I come to Mexico. [[Know]] I know what you look like. [[YouFar]] You will be/are far away. [[YouGo]] You go to Mexico. a) YouGo > YouFar b) YouFar > -Hear(*) c) -Hear > -Know(*)(+) d) (ICome ^ -Know) > -Find e) (YouGo ^ ICome) > -Find(#) (*) Note that here I am assuming that "A when B" can be represented as "B > A". This is not always the case, since "when" has temporal overtones that cannot be captured in propositional logic. But making this assumption allows us to represent the argument in a way that it can be shown to be valid. (+) Here, I am assuming that, at the very least, "I forget what you look like" > "I don't know what you look like". (#) Similarly, here I am assuming that, at the very least, "I follow you" > "I follow you to Mexico", and also that "I follow you to Mexico" > "I come to Mexico". Grading of representation: 0 = missing 1 = incorrect translation 2 = correct, but not for validity 3 = translation that enables proof of validity Proof: 1. YouGo > YouFar : assumption 2. YouFar > -Hear : assumption 3. -Hear > -Know : assumption 4. (ICome ^ -Know) > -Find : assumption {Show: (YouGo ^ ICome) > -Find Strategy: Assume antecedent & show consequent; Then use >Intro} * 5. YouGo ^ ICome : temporary assumption for >Intro {Show: -Find Strategy: Show (ICome ^ -Know) & use >Elim; Note: Other strategies are possible; e.g., assume Find, and use -Intro; but direct proofs are generally preferable to indirect proofs by refutation (unless, of course, you're doing automated theorem proving :-)} * 6. YouGo > YouFar :send 1 * 7. YouGo :5, ^Elim * 8. YouFar :6,7,>Elim * 9. YouFar > -Hear :send 2 *10. -Hear :9,8,>Elim *11. -Hear > -Know :send 3 *12. -Know :11,10,>Elim *13. ICome :5,^Elim *14. Icome ^ -Know :13,12,^Intro *15. (ICome ^ -Know) > -Find :send 4 *16. -Find :15,14,>Elim *17. (YouGo ^ ICome) > -Find :5,17,>Intgro 18. (YouGo ^ ICome) > -Find :return 17 Grading: 0 = missing 1 = seriously illegal 2 = correct up to a point, then errs 3 = syntactically valid proof Total = 6 points ======================================================================== Grand total = 15 points 463 both 563 A 15 A- 14 B+ 13 B- 12 C+ 11 C 9-10 6-10 C- 8 D+ 6-7 D 4-5 F 0-3 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 14:18:38 2005 for ; Mon, 28 Feb 2005 14:18:37 -0500 (EST) for ; Mon, 28 Feb 2005 14:18:37 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 14:17:28 -0500 19:17:28 -0000 j1SJHSx3023720 for ; Mon, 28 Feb 2005 14:17:28 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Feb 2005 14:17:28 -0500 (EST) Date: Mon, 28 Feb 2005 14:17:28 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: TRANSLATIONS FROM ENGLISH TO F.O.L. To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------- Subject: TRANSLATIONS FROM ENGLISH TO F.O.L. ------------------------------------------------------------------------- To help you study for the midterm exam, here are some sample translations that are relatively straightforward. The English sentences are from a logic textbook by Lewis Carroll (yes, the author of "Alice in Wonderland"!). First, here are some sentences to translate; choose an appropriate syntax and semantics: 1. No kitten who loves fish is unteachable. 2. No kitten without a tail will play with a gorilla. 3. Kittens with whiskers always love fish. 4. No teachable kitten has green eyes. 5. No kittens have tails unless they have whiskers. Therefore: 6. No kitten with green eyes will play with a gorilla. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Now for the translations. Let [[K(x)]] = [[x]] is a kitten [[L(x)]] = [[x]] loves fish [[T(x)]] = [[x]] is teachable [[H(x)]] = [[x]] has a tail [[P(x)]] = [[x]] plays with a gorilla [[W(x)]] = [[x]] has whiskers [[G(x)]] = [[x]] has green eyes I will use "E" for the existential quantifier and "A" for the universal quantifier; I will use "-" for negation, "^" for conjunction, "v" for inclusive disjunction, ">" for the material conditional, and "==" for the biconditional. I will also omit parentheses where they are not needed (including around the arguments of 1-place predicates). 1. -Ex[Kx ^ Lx ^ -Tx] 2. -Ex[Kx ^ -Hx ^ Px] 3. Ax[(Kx ^ Wx) > Lx] 4. -Ex[Kx ^ Tx ^ Gx] 5. Ax[Kx > (Wx == Hx)] Therefore: 6. -Ex[Kx ^ Gx ^ Px] Some comments: A. The test of whether these representations are OK is if you can prove 6 from 1-5, either semantically or syntactically. We'll revisit this later. B. Translating "unless" is tricky. Construct a truth table: P Q P unless Q T T F T F T F T T F F F According to this truth table, "P unless Q" is equivalent to "P xor Q" (or to: -(P == Q)). There is also a "weaker" reading of "unless", which makes the conditional only go in one direction. See which one is actually needed in the proof of 6 from 1-5 to see what that weaker reading is. (And I'll post a followup after you've had a chance to think about it.) From rapaport@cse.Buffalo.EDU Mon Feb 28 14:25:48 2005 Mon, 28 Feb 2005 14:25:48 -0500 (EST) for ; Mon, 28 Feb 2005 14:25:47 -0500 (EST) Mon, 28 Feb 2005 14:25:47 -0500 (EST) Mon, 28 Feb 2005 14:25:47 -0500 (EST) Date: Mon, 28 Feb 2005 14:25:47 -0500 (EST) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu Subject: CSE 4/563: EINSTEIN AND GOEDEL Cc: snerg@cse.Buffalo.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: EINSTEIN AND GOEDEL ------------------------------------------------------------------------ There is a *wonderful* article in the current New Yorker magazine about the friendship between Einstein and Goedel: Holt, Jim (2005), "Time Bandits", The New Yorker (Feb. 28): 80-85. http://www.newyorker.com/printable/?critics/050228crat_atlarge Note: If and when this website disappears, a copy will be available at: http://www.cse.buffalo.edu/~rapaport/563S05/Private/einstein-goedel-nykr.html Reminder: Login = CSE563 Password = Spring2005 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 17:05:18 2005 for ; Mon, 28 Feb 2005 17:05:17 -0500 (EST) for ; Mon, 28 Feb 2005 17:05:17 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 17:04:08 -0500 22:04:08 -0000 j1SM46x3025373 for ; Mon, 28 Feb 2005 17:04:06 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Feb 2005 17:04:06 -0500 (EST) Date: Mon, 28 Feb 2005 17:04:06 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: STUDY QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A student writes: | Is there any text book that is available in our library that has some | practice questions for the following topics that would help us in our | exam preparation: | | 1) Conversion to Propositional Logic | 2) Syntactic Inferences - proof theory | 3) Conversion to FOL | | Or if there are any resources available online to check for sample | exercises, that would be great! There are plenty! Any good logic text will have all of the above, though, as I've said, the notation and rules of inference may differ from what we are using. I've also put several items on the course website, either in PDF or at least cited by name. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Feb 28 21:40:19 2005 for ; Mon, 28 Feb 2005 21:40:19 -0500 (EST) for ; Mon, 28 Feb 2005 21:40:19 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Feb 2005 21:37:14 -0500 02:37:14 -0000 j212bEtH028895 for ; Mon, 28 Feb 2005 21:37:14 -0500 (EST) 21:37:14 -0500 (EST) Date: Mon, 28 Feb 2005 21:37:14 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Some Practice Derivations To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, Here are some syntactic derivations for extra practice 1. (C > D) > C |- (C > D) > D 2. {-A, B, B > (A v C)} |- C 3. {A ^ B, -(A ^ -C)} |- C 4. {-(A ^ -B), -B} |- -A 5. {A > (B > C), A > B} |- A > C Solutions below... I'll try to post some translation problems tomorrow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1| (C > D) > C |- (C > D) > D - Proof: 1.(C > D) > C :premise *2.(C > D) :temp assumption *3.(C > D) > C :send 1 *4.C :>Elim 2,3 *5.D :>Elim 4,2 *6.(C > D) > D :>Intro 2,5 7.(C > D) > D :return 6 2| {-A, B, B > (A v C)} |- C - Proof: 1.-A :premise 2.B :premise 3.B > (A v C) :premise 4.A v C :>Elim 2,3 5.C :vElim 4,1 3| {A ^ B, -(A ^ -C)} |- C - Proof: 1.A ^ B :premise 2.-(A ^ -C) :premise *3.-C :temp assumption *4.A ^ B :send 1 *5.A :^Elim 4 *6.A ^ -C :^Intro 5,3 *7.-(A ^ -C) :send 2 *8.C :-Elim 3,6,7 9.C :return 8 4| {-(A ^ -B), -B} |- -A - Proof: 1.-(A ^ -B) :premise 2.-B :premise *3.A :temp assumption *4.-B :send 2 *5.A ^ -B :^Intro 3,4 *6.-(A ^ -B) :send 1 *7.-A :-Intro 3,5,6 8.-A :return 7 5| {A > (B > C), A > B} |- A > C - Proof: 1.A > (B > C) :premise 2.A > B :premise *3. A :temp assumption *4. A > B :send 2 *5. B :>Elim 3,4 *6. A > (B > C) :send 1 *7. B > C :>Elim 3,6 *8. C :>Elim 5,7 *9. A > C :>Intro 3,8 10. A > C :return 9 Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 1 13:30:31 2005 for ; Tue, 1 Mar 2005 13:30:30 -0500 (EST) for ; Tue, 1 Mar 2005 13:30:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 1 Mar 2005 13:28:24 -0500 18:28:24 -0000 j21ISOx3001381 for ; Tue, 1 Mar 2005 13:28:24 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 1 Mar 2005 13:28:24 -0500 (EST) Date: Tue, 1 Mar 2005 13:28:24 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: "UNLESS" To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: "UNLESS" ------------------------------------------------------------------------ Some of you have indicated that you are still a bit confused about how to represent/translate "unless" into a language for propositional logic. Join the crowd :-) In fact, I used to think I understood it perfectly, and have now thought about it so much that I no longer have any intuitions left about what it ought to mean. However, I have found a webpage that does a pretty good job: Rice, Hugh (compiler?) (2002), "Introduction to Logic" http://logic.philosophy.ox.ac.uk/ especially: "Introduction to Logic: Level Two Tutorials" http://logic.philosophy.ox.ac.uk/tutorial2/Tut2-b-02.htm To summarize this page (which, unfortunately, uses unreadable connectives :-( "You will not win the lottery jackpot unless you buy a ticket" is best understandable as: If you win the lottery, then you must have bought a ticket. Or: You win the lottery > You bought a ticket. But note that the original sentence had the form "-P unless Q"; its best translation is as: P > Q. Note that this is logically equivalent to: -P v Q. But that means that the best translation of: A unless B is: A v B (For details, read the web page cited above.) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 1 18:42:11 2005 for ; Tue, 1 Mar 2005 18:42:11 -0500 (EST) for ; Tue, 1 Mar 2005 18:42:10 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 1 Mar 2005 18:41:43 -0500 23:41:43 -0000 X-Originating-IP: 67.23.179.193 Date: Tue, 1 Mar 2005 18:41:42 -0500 Reply-To: Paula Chesley From: Paula Chesley To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU This message is in MIME format. ---MOQ1109720502ee4e892280cb5c59e0665b145f4a45dd Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit Hi, I had sent this email to Dr. Rapaport, and he suggested I send it to the listserv. In perusing the blog located at http://pasfolle.blogspot.com/ and scrolling down a bit, you will come to the images I've attached here. **What a real-life application of "Three boys invited four girls"!** As we were studying these types of quantifier resolution problems, I thought nobody would be so demented as to actually naturally produce these sentences (because they seem really unnatural to me), but of course I've been proven wrong. The saddest thing is that it appears that the 7-year-old child got the problem wrong, when really the answer is completely plausible, and it's the methodology of the teacher that should be in question here if anything (wait-actually I contacted the blog writer, who has since told me that she doesn't know if her child got this question wrong, because the exercises usually aren't handed back to the pupils). Which only strengthens my belief that all educators really must take at least one linguistics class in their career (no ambiguities there, I hope!). :) Paula *************** "Tiger! Tiger! Burning bright in the forests of the night." Blake wrote that. Apparently the tiger was on fire. Maybe his tail got struck by lightning or something. Flammable felines--what a weird subject for poetry. ~Calvin, Calvin and Hobbes Bill Watterson ---MOQ1109720502ee4e892280cb5c59e0665b145f4a45dd Content-Type: image/jpeg; name="24HeartsProblem.jpg" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="24HeartsProblem.jpg" /9j/4AAQSkZJRgABAQEAYABgAAD/7QEEUGhvdG9zaG9wIDMuMAA4QklNBAQAAAAAAOccAlAAC lBpY2FzYSAyLjAcAnUAvzxoZWxsb3N0YW1wPgogPGdpZD5mMDUxMC0xLWZmZmZmZmY5LTY8L2 dpZD4KIDxtZDU+Y2I2MTdlMzgtOGVmMDJiNmEtYjZmZTE4YWQtZTMxYTdmOGM8L21kNT4KIDx vcmlnV2lkdGg+NTk3PC9vcmlnV2lkdGg+CiA8b3JpZ0hlaWdodD43NDwvb3JpZ0hlaWdodD4K IDxvcmlnU2l6ZT4xNzc5OTwvb3JpZ1NpemU+CjwvaGVsbG9zdGFtcD4KHAJ2AA88cGljYXNhc 3RhbXAvPgoA/+AAEEpGSUYAAQEAAAEAAQAA/9sAQwAGBAUGBQQGBgUGBwcGCAoQCgoJCQoUDg 8MEBcUGBgXFBYWGh0lHxobIxwWFiAsICMmJykqKRkfLTAtKDAlKCko/9sAQwEHBwcKCAoTCgo TKBoWGigoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgo /8QBogAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoLEAACAQMDAgQDBQUEBAAAAX0BAgMAB BEFEiExQQYTUWEHInEUMoGRoQgjQrHBFVLR8CQzYnKCCQoWFxgZGiUmJygpKjQ1Njc4OTpDRE VGR0hJSlNUVVZXWFlaY2RlZmdoaWpzdHV2d3h5eoOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKm qsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4eLj5OXm5+jp6vHy8/T19vf4+foBAAMBAQEB AQEBAQEAAAAAAAABAgMEBQYHCAkKCxEAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhc RMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVl dYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm 6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/8AAEQgAMgGQAwEiAAIRAQMR Af/aAAwDAQACEQMRAD8A9sDAfSjfzwRzUOeOK811Sy1238X6fpkfi7UxBfRTygmKIlCmCFHHI wf0r0JOx87CCk3d2PTmeoJMHnrXIjxYunw3NtdC6u5bK9g057h9imaSQD5sDgYyM1TsPiJY32 tvp8NndNHtmaO4BRg4jzuO0HcoODgkDNLnRapT6HcZGKCBmuDsPiELnSptSbSLtLban2crNHJ 57uwVY/lPysc5we1Z0Hiq6s/Eut32s6fd2vk2ttDFYCUSmSR3bbtx8uT69sUnUQ1QkenIcVOp yOf1rgLHx6rTpHqWl3OnkXq2MzSyIyxSMm5ckdj09jWvYeJbjVbZrjSNOM0KX/2QvLKEDxg4e VfXB4A70+dEujJbo6k4x7mgHB7/AJVy/wASppYPBWpSW7zpIAgDQMVfmRQdpHOcE9Kw9EvLfQ 4NR1A2Pii3VNsSW2oTGUTuzYTy9zH5icDqMZ5ocgVK6uehkjHYUxvQYFcta+L/ADp4LW506e1 vXvlspYHdWMRaNpA+RwRgdqguPGji4FvZ6Rd3k7XNxbIsciAZhxuYk8Ac/wCc0cyD2UjrGU7W AOMg1U0908pYGwJ4lCtGevHfHcH1qrHrcX/CNLrN5DLZxC38+SKQfPH/ALJHr/jXJ+KtXuZ9A hu7/Rr2w8q9tWVd6yPIjPkquw5zjgqe570OVhxg3ozvobWCOdnjjjWRvvMByaSz/eardvxtjV IB9eWb+Y/KuPi1Vbu3tY/+Jxpcp1NNPkthMjFWILYJ5+Ur1wc0eGdVfStNngt7S51C8udTvFg tkYBtqsMksxwFAx19aOfuP2bSZ6A/Ce1QZ3c4GO1cDpXjC4trNy+male3V/d300dsrKWhWNlB RiTgAc9P61oS+O7YanZ2i2NywnMUbtuUNG8ihguwnc20EFsDA/CmpoTpTR1ZXrmnEciuU0nxi NR1dbT+y7qKFria1F0zqUaSMEkAdTkDOex4qfxD4vs9CmuUu4pf3MUUoKkfP5khQAZ7gjn2oU 9NyfZyvY6bbxRgY5rlbnxlGkk8Fjp9zeXKTTRiOJlUFIgN8hYnAXJwM8kiqdz49EbA2+kXk8f 2GPUJH3qoiiYnO7PcYHTrQ5oPZTZ2bDA4/KmsmRXKrr+rP42FgNNf7A1uHVvOj+7vx5vrjBxt 61J411jU9LfT/wCzbF7hJbmOORvMQA5JHl4bufXtS5uoeylex0nGDTRwTjr7Vzc/iwQ3627ad c+UrxQ3M4ZSkEkn3UPOW6jJHTIqbxX4h/sKG2C26zXF07Rxq8qwoCFydztwBT5tLi9m72OiHQ U/HavOG8T3ema7ezXemXUn+hWs1xHHIpS2HzBjknBPI4HJxW5J4t26wlrDp881mbtbFrwSKqi YjOAp5YDufWp50U6MjrR7UmM5xxXFW3jmRrVr270S7ttP8mWVblpFYOUbbjA6biQAT159K0fC niZ9flvEeyMH2YqPMSdZo3J7B14LDuBnFPmTJdOSV2dHgEcCoZ41lR0kG5HG0j1B61w/jdHn1 TytO1HU5dadFNrbW8vlxWoyMyyY42nn72c9AKx/iD4ljkvxpqaqloLCSGS4xJseeUuvyAf3VU lifoKXMXGlc9P24UD8BTHTKn36kVz3xBuE/wCEG1O4gJkQxBkMTfeG5cbTVJ/GclrZ6g+p6U9 nPapE8cL3CEyCQ7Uy3RDkc56U3KwlTbV0bkyX1tzGVu4+hRsJIPx6H8cVS1S+zYSxy2l1DIwG 0NHuXIYHGVyKzrHxjc3dirppW+7muvstssdx+6nO0szLIVHygA5OPpmo18X6jOlrFYaH5t9OL jMZugqRmF9jZbbyCenHpU81ylB32O3Qce/pVhF/KuDh1jxG/i/TYBp1vBaXFmZWhlufmHzJuY jbncuduO4q54X8V3uveIJYI7azgsIfNV42uM3UZU4VnjIG1Tz0z25pqa2JlSdrnbBe38qH645 rn/iE7ReDNWdXdWEBO5M5HzL0x361ga94kGoSaPa2lnq8Mcmo24a5ntmhjIB5GScnOPSqcrEw puSujt3eos5PeuQj8Z3E0iyrpDiwuPOFpP8AaFzM0as2CmMqG2HB5qKHx5bzSNFb2kjy+VayI u/Adpyvyg4/hDAmlzor2Uux2wXJp4UVz/ifW7nSJ9NhstPa/lvJnhEayBMYQsDk8Acck9BWVY eO2a0e61HS3tYGs2vIStwkrSqrqhXAAwdzKBz37UOaQvZSauduF+Ye9PC+lcPc+NNQs4rpbnQ St7BcW0X2aO6WQsJt20BgANwxyOgq5pvjPN2bfWbIWDqZ43kWYSRiSIByoOBnKHdn1BFHtEHs pHXdMYFOz9K5+TxBJH4Ok157GRStsbkWzP8ANt6gZxwduD0qlb+MoLnVJLK1tXkZNQisVIcAM GQs0g4+6u0j3NNyRKpuWx1metKpz/jXmcd39qfSJ7d7oRS6Zqku2aYyNkFR146c49BXbeEtx8 MaQzMzE2kRJJyT8goUruw5U+VXLmcLwOa4e/j1y71+w1M+HyJLJZURBqcIDh8DJ+XPau5zj0q NpFEmzzF3YzsyM49cUNXCErdDy7UfDviC91qa8GnSQ2s93FeTWQvrdkeSMAZ3FdwyBUejeGfE miyv/Z1tLBbEuRCl3aqctnBd/L3PgnIB9K9T85WcgOpZeoB5H1rO1DWYrW40mOMLMt/cGASI4 ITEbPu9/u4/Gp5Ym6rTeiR5rZ+CtZ/4mD3doftV0I/9IS9hh8tkYMrhY0wWyOpq3J4M1+druW 6nFxe3DRMty94A0RiJKFQIsdzmvUlZZEVkYMrcgjkVk69rMGm6XqNzBJDcXFnC0rW/mjPHY9x +VJwiT7abZ5xrHh3xFbeH9XtZraLVJtXmj3u05Zo3OB5gVYwABgHPbFdGmi6wukWGljSNK+y2 LxvFjUZVO5OjHCc85PPXNdTNqSrpk1zAqzyxIGaBJlVskD5SScA89/SotO1nz73V4biNLeGwk jTzHkGG3RhiT2GM4pKEV1G6k30JNb0uXXNCFncSJbTu0UjmPLqpVlYgHg44xn3qfxDpSazpz2 rSvbyb1ljmjALRurblYA8HB7GodF1ldUvdUhhCGCzkSNZkcMJcoGPTjjOK0jcRGXyhLGZcZ2b hux6461poZe8jl5vBztbbm1e5OqfbFvPt/kpneqFFGz7u0KSMVLonhU6bcW9xLqU13NFJcyl3 iVC7Tbckgccbe3rW819a5mBuYP3PMn7wfu/97nj8axB4ttrnTFutJtbnUnMxh+z220yAgsCzZ OFX5TyfUUtCrzaNTWdOi1fS7qxmd447hCpdPvKcggj6EZrLTw3e3ESLq2tTXpS5iuV/0eOIDY c4wvr3NaegatBrdg1xbxzQNHI0MsM67ZInXqrD15H1zVN/Ea/2jNa2OmanfRwS+RPcW8a+VE/ dcsQTjvjpT03ElNaIQeFYm1J7trl+dSj1JY1UYDomzaT6HrUJ8HyQwQHTtVmtbyCe4lS48lHy Jm3MhRuCOF59q1dJ1YXt1qMTKsJtbprZSWGZMKrZH59K1RMhYqXTeBkjdyB6/T3o0YnKaOb0T wo2mXllO2oyXT24uizSRBWkadlYn5eBgqenrVa78JSDX7vU9O1SWxN26vOsdtG0hIAGFkYblB xyPrXXGWMrnzFxgHqO/So3Yb8E8+n44oshc8jmbbwwtqNP8u6c/Y72W75QfP5m7K/+Pdfam6/ 4Stta1SK8uJnXy7aS38sD5WLfdY+65JHvXRmWMFf3ifOcL8w+b6etIJYs/wCsT72z7w+96fX2 o90alPc5SLwXLb2mnx2Orz29xb272803kpIbgO25iQ3Rs96mtvBqQ6VeWT3skhudPTTzIY1G1 UDgN9fn6e1dOs8UjmNJEZ8biqsCQPXFPE0RIG9c52feHX0+vtRaIOc9jKi0dY9Zi1AzszpZi0 2bQAfm3bvrRremDVI7ZDIyeRcxXOVAO7Yc4/GtIzwjb+9jAc4X5hyfb1psk8CyLGZYw7ttClh nPp9aNBe82czd+FvP1aS5+3zLZyzJcy2gRdryoBht3XHygkd8CrHibRJ9ZiSGPUXtIsMsiCCO USAj0ccEc4Iq14f1mPV9KtLt1SCS53lImcZO1iDj16ZrSeaFJER5EV5PuKxALfQd6NBtyuc6/ hK2awv7YXE4W7s4rTc2CUVFIUj1POahk8Gub9JotYvI7Zbtb5bURRlfNByTkjdg88dia6eK+t JJfKjuIGl5+RZATwcHjPaqWq6/p+l31naXdxHHNcsyrudRtwu7LZPAOMA+tJqI1KeyK/8AwjV s3hX+wZ5JJLYxmMvgBuWLA+nBNGh6HPpcxefVLu8ATykjdUjjjGc5CIAM+9S6PrSX819FL5UL wXkltGPMyZQgB3Afj0FXTeWxuzbG4h+0gbvK8wb8eu3rRoJ860MS88MtNq13f22s6pYvcBN6W zoFOwYB5Un8Kv6zpVvqdgbS4ZgGaNjIoG47WDdce361FqmtRabqun2tyYo4blZWaaWQIE2AED n1zVq41C0t4Fnubu3igfG2R5VCtn0JODTVgvLRkes6bDq2mT2VwXSKbAbyyAcAg4H5Vm6t4Xs 9UuLqa4kmU3MMcRCEDaY33owz3B9eK07rVLC18n7Ve20PnH90JJVXzP8AdyefwqjZ6/E1zqqX 7wW8Nndx2yOz43lkVh17ktjijRguZbFNvCRcJJLrOqPfxuHiumdcxYBGFTbtAIJB4q1oXhSHS ruG4F7d3EkYmAM5Uk+ays2SAO65/Gtm7ureyiM15PFbxA4MkrhFB9MmsXxP4kGnJYw2EunteX j4Rrq4CRom0sXYg5xxgepNJ8qKTnI1ZtIjn1m21EySCWGCSBVX7uHIyfqMfzrNsvCTW91JdS6 zqVzefZ3toJ5iha3VjyVwo3Hgct6dK1NV1e20bRnv9RliREj3H5wodtudqk9SecVmL4us5LvS /IuLR9Ou4JpZZzOuImQKcE52/wAWDnoaHa4lz7I1dY0ldV0V9NuZ5QsioGlXG4lWVs+nO39TS 6xpkWp/ZvOZ1NvcpcoU7spyAfbmnPqWbyzjt1gmt7iN5POE6ggADBC9WBz1HA71Fp2vaVqt3c 22m6hbXU1tjzVhkDbc9MkcU00yfeRj6d4RtbG9WUXN3Nbw+Yba0lcGK3353bRjPOSOScA1W0z wLpthcaTLDJOW05pGTc3+sL8jd67eMfQV0Y1TTpLx7VL61N0rbDCJV3hsZxtznOOaVdQtBfGz +125uxyYfNXzMf7vWiyHeoMvNKivL7T7qRnD2UjSIB0JZCpz+BrGPgeyksLS0muLh47aCa3BB CsyyMG5x3BAI+ldFbahZXDypDd28kkOfMVZVJTHXdg8fjUmn39nqERksLq3ukU7S0EiuAfQ4N GjJ5pJHPWvg4Jfm8u9Wv7ydpYJmabZjdDu24AUYGGI/GqniXwsl40VksElxBf6qt5cuw4gVU5 xj+9tC/8AAjXb5+agjBp8qD2ktyve2UV7YXFpMMRTRtGwHoRg1gaN4M0/Sb/T7q2knZrK1+yq jnIfknzG9X5bn/arqBjrSkgDvTshKbWxzkHhKzhjtUSa4xb21xapkg5WYgsScdRjitnT7eOw0 +3s4yWSCJYlLdSFAHP5VYZvcCo854oE5N6MqkDIz0zXkNjpt6niW5uLiyv5NdS4uTC0NltUhg 4SSS4Y4ZACuEXAGOmc17Js/KnBPpUuNy4VHA8UtrFvssw03R9Whun0sRX7G2ZJJpRNG043t99 ym/B6Y4Brcl0TTdcOjQ6XoF3Z6QupGSdWtzbrIggYbtuchSdqHgZ/WvUREN38qd5f/wBfNLkR brvexxvgWznstAu7I27WxgvbpIEZcLs8wlCB/dwRjHauUtdKlm8KX+njwvfR60bKaOa+eJAZZ WPzbZM5bceR6Yr13y/8KVYwD2p8pPtXe55Pq/g+e0i1Ox0PSyLe40y3T92o/eyrcKW3Hu23Jy ava3ol2b7UrmXS57+x/teKd7VACbiBYNoKqThgGwcHrivTXAHpULkYPTn1o5EP20nuch4BsPs w1mVdIl0iC4vPNht5UVfl2KM7VJA5B4rjItOuotW0KCXR5rXWGv5pZ9TmUfv3EchXa+SzLypx wAABXr3Uc498Vh6Z4X03TtR+2xLcy3W1lV7i4ebyw3XbuPGfapcOhUaurkzgvAWgmHULddW0K /e7MJt7mSa1iit1DcuWbJM5Yjqc9c4FbOhafPovg68OnacLPU5rqWJHSzDOFachGKjqoU5GeA BnpXoAxnjrUiimoJESqtmVoGixaLYtBHLLcSSSNLNPMcvNIerH8ug6CuA1SJru7uX0TRde0zx JJNv3KzLbbtw/eOc+WykcnAyfSvVsZPOKAn0puNxRqNNs8t1Tw0yXOr60dPma/XWoZbd9pLeS rRhmjHYHnJHUD0qKx0h7nxGsUul3iawL+4kvdQdCI3tWDqFD5wylWUBR0wa9ZwCaQjjrml7NF +2fU8Y0PSNalutHF/a3CxXqxWc6lflhitHDAnj+PDEZ9RXXfEKK9spdP1bRio1Fn+wbGOBKs3 A/FWww9ga7g8g1SvtOtry6srifeXtJDLEN2BuKlckd8A8UcmgOs21Jnlus+HPsd0dOj028u7g W1vDpV2iFkt3BJkctnCNu+Yk9QcVZuTLHrCaW+m6h5i64975/2cmEowYh9449Pyr1IjjjGKTA yB+NHs0CreRx3w/8L2Ol6Rpl7HZtBqD2YjmZsgsScncD3B6eg4rm/EWm6xb+INTvbK1nkgsbx NUhWMf8fEkiqjKP90Kx49a9YC+ppcd880+RNWJVVp3PH9W8PSWUKWE2k3eoXLackGnzRJuS3u MsZGJz8h3ENu9Biuk0nwxa3mu6rfalZs13FfxSRTnIJ2RocqfQtnOOvQ13m0YwKaRihQQ3Wke Z+A9D1DSr7T7jUYJbgTwSIjOmDYEOzbP91gR83XOB0p3inSbm68S3edLubq5uBbixvVUGO1CN lyWzlD1PHXOK9KwKTAyT1o5UL2rvc820rw89lq1jqI090vJdauXnk8v5hCyuF5/uHj8TW/4p0 WK+1vQrhtOhudk7LPI0SttTymxuJHQNjHvXVEcccUm0AUcqB1ZN3PNpPDDxTX+qLpZbUG1yKe N9oL/ZxIgJXngbS2cdap6Pod7Z+JxPqGlX93eJeySx3CRxRwAMxxI0ud7kKR8px0xivVTj0FR PgH3+lLkQ1WdrHH+MbVn1fSLo6JLq9vAJg8ccaSFCwXacOQOxrnW0S9tRG1xoMt3by21wlvax 7HFk8khIyGOB8pAyM45FeoHGeeaYeaHC441HE8206yvrJ4JL/wAMz6jvsLe3jXERFuUBDq288 Zb5sjrxUs3h3Ux4ju9c+zvcCO/V47BypR4zGiGVfSQdQT2Br0YAEDgU4AZo5EDqs5LxlZTy32 lXa6U2r2ls8vm2iBCSzLhXw3BxyPbOazLDw2YbjwjFfabBJJEtwLgmISLGpUlELY6AtgduDiv QtnJqVFweM0+REqq0rIz9esEvdEvbbyEmZ7eQJGVB+bYQuM9OcVzEXhf7Vc+D/t+mQtb2Nmxu EZF2rMY0ABXuchvXpXfBemaCg7cU3G5MajitDzGx8P6jBY6clxpc00drZX9s8Mcqxs2+QbApz xuXoe1anw+06/smvVurGe2skjjjtftaQicAZ3LmLgoOME859a7sqM8004wfTFLlKdV2seUXOl y6j4i8Q21lowad9Uhf+0zsAtwojY/7WeOg4O6o10HVDNaQQaHcQ30d1dyT6mdm2QypIqSbgd5 xuX6V6jZ2VtbT3UsCBJLqTzZjz87YAz7cAD8KmuYIriCWGVcxyKVYA4yD15HNLkK9s9keQSaN eXryabZaKNKuoNDMQhYx7pWMseSSvY7GALHnJ6V1XhOKfSPtcx0DUla5eISyy+SJJGyVGI4yF VEU9f510eiaDpmiCX+zbcRNLje7SPIzY6AsxJwPStTt2xTULahOrzKwvIP0p2eOc0wEZ4NG7J qjAex54xzTCc8U0nmmscHrQASHjimA80NSD9KAJk/rTlOXOfSmp0/GnL98/SgB56j6U5f6Uh6 /hSr/AEoAU9fwpknSnnr+FMl6UAR9hTX6Gndh9Ka/Q0ARp94/SlH+soT7x+lH/LU0AxyfeNSf xmo0+81SfxmgEOIo/hoo/hoAD938KDQfu/hQaAI+1NHX8ad2po6/jQDB/wCtD/dWh/60P91fw pjew8mnjqPpUZqQfeH0pCE70HvR3FB70ANakf8ArStSP/WgAPQUN0oPQUN0oAYf4aZL0p5/hp kv3aAKzdTTm7Ux+ppzdqCh6/dNSr0NRL901KvQ/SgBw+7VhPumq4+5VhPumgkkPSg/dFB6UH7 ooAa33qa3WnN96mt1FAESdTUg/rUadTUg/rQCGDqKcvQ00dRTl70wD+IU1+tOP3hTX60gIz96 hulB+9Q3SgBjU9e1MNPXtQAA/9k= ---MOQ1109720502ee4e892280cb5c59e0665b145f4a45dd Content-Type: image/jpeg; name="24HeartsAnswer.jpg" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="24HeartsAnswer.jpg" /9j/4AAQSkZJRgABAQEAYABgAAD/7QEEUGhvdG9zaG9wIDMuMAA4QklNBAQAAAAAAOgcAlAAC lBpY2FzYSAyLjAcAnUAwDxoZWxsb3N0YW1wPgogPGdpZD5mMDUxMC0xLWZmZmZmZmY4LTY8L2 dpZD4KIDxtZDU+M2UzODQwZDItODQwMThmMmEtZmNhNzZlZjktZjA3NjE4MDk8L21kNT4KIDx vcmlnV2lkdGg+NTY4PC9vcmlnV2lkdGg+CiA8b3JpZ0hlaWdodD4xMjA8L29yaWdIZWlnaHQ+ CiA8b3JpZ1NpemU+MTYwNTg8L29yaWdTaXplPgo8L2hlbGxvc3RhbXA+ChwCdgAPPHBpY2FzY XN0YW1wLz4K/+AAEEpGSUYAAQEAAAEAAQAA/9sAQwAGBAUGBQQGBgUGBwcGCAoQCgoJCQoUDg 8MEBcUGBgXFBYWGh0lHxobIxwWFiAsICMmJykqKRkfLTAtKDAlKCko/9sAQwEHBwcKCAoTCgo TKBoWGigoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgo /8QBogAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoLEAACAQMDAgQDBQUEBAAAAX0BAgMAB BEFEiExQQYTUWEHInEUMoGRoQgjQrHBFVLR8CQzYnKCCQoWFxgZGiUmJygpKjQ1Njc4OTpDRE VGR0hJSlNUVVZXWFlaY2RlZmdoaWpzdHV2d3h5eoOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKm qsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4eLj5OXm5+jp6vHy8/T19vf4+foBAAMBAQEB AQEBAQEAAAAAAAABAgMEBQYHCAkKCxEAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhc RMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVl dYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm 6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/8AAEQgAVQGQAwEiAAIRAQMR Af/aAAwDAQACEQMRAD8A9iK896TB98e1SH7wpGKopZyFUAkknAA+tegfOERX8qcoxTIJ4bpN9 tNHLH/eRgRUmMH39O9A7McvAHSnrk/jTPTAzSeagfYXXdjO3POKLiLCnPWngkjIqlHdwSSbEm jLj+EHJqw0yQoWmdUX1Y4FFwsWMZxUbr35oilVxmJlfPdTmnMN2exp3Bplc9TUTdasPHg84NM IOeR+OaAIjjFNYc1LtwTgZpGBHrzSGRYpuCODUpXDc008knmgCPqTignA5p5HJ7CmtjPNACA8 UoPrTSOvWlXGepoAUYBpw/D60AU7HFAC8ng1ZiOPvc1VJwfrUsZ7Hp6UATsBjOePSoHYDuakn yV4JAx0qmy565oAJHAB5qItnp0oK5pVUE0AAxSnrxTsemcUEE0AJTxxkcYpMdM0pzzlaAA8dK dkbvekA/Ol+vegCRGw1SK1QLz0p6nGPWgRYVutPD9KrhuaeDzQBYIDjnmqzrt6LUinB4p+M9a YimRjmkI7mpWHzYxTHHqKBjc80o9qZ3+WjJzzSAmVuccVIOlQrgn0qVW9KYD87SMc0pOSOlIv Xp0pRmgTY0nnHSmMPWpXGfrTCAP/ANVMEQHqMjNY2tE3d3aacSRBOGkmx/EiY+T8SRn2FbzqT XM6/prNqtlfw3c1qwzBIybWCg/dJVgQPmAGeOvtUmkLXJtXdrCSO4shGskrLbsp4U7jhWI9j+ nFQXVrBpVsdQJeS6jIZ5STulycFf16duKr63YXZNgJ9UlcfaowMQRqR156HpzVHxNbs8a29rc XMs0bxzXDyzMyqu4YyBgcnnAxgDNZ7GqSdtTauIljsnu9W3TyZBMYYhIwSBhR7Z69TUC7G0Wz k63E86q0h+8csVPP0FVw2pzTLYx29tttWWSYm5dlY9VXJXPbdj0x61TvnvzG2xYIoLF1AG7zF Zy4z83B4B9Pxp3v0HynYzxRG3MaqFUD5do+6QODVPTLyDUJJ5V+bySIuV4zgE/XrVHVNXu7Cz ln1C2byo1LtLaPuXpwCDhhkkdM9ag0KeODQoLexnSS9k+/jqJG+ZmIPIA/pii+pny2V2TRv/a WtXVumUsowGkEZ2+Y3QDj6Mfyq/cu+k+WbYvIkrCJYXYkBj90gnnGRz7Vi+Y+h+JcIkklnPao ZSoJaMoSu/Hcc84+taV1qVrfa1pcUF3CbaJXu5G3jaeNqDPrkk/hQipLVFi9l1CxtnvJ7yAxQ /O6CDG5QeRnOenepY9QuJIvOSyf7ORlSXAdl9duPxxnNUPF+oW8+nRWcF1A/wBqlSN9kinCZB bv36fjVnxBrSWdmws2ilmYqi4+ZRlsc49snHtVX1JtdbDrnVgAfs0ElwyR+a+3gBcZ6nvjsKv xTJLDHKmdrqGXI5wRmqdjF9k0Xc4OTEZHY9+Cefw/lWfps9zOkNpbOIEtreIyOybmJZcgAHgD HPPNK4ctzdJyTmmFxkccVnG/uIL37G8X2mVk8xWiAX5c4+bJ4/Ckh1cSzzw/Z5DNE2CkZDduu eAMU7k8poEjPP8AKms6+9Vn1GGME3BMTAhSjjLAnp0zn8Kswyx3ESyQsrI3IK807isIp69acO uf1pycnGc0/KeZs3pv67cjP5UC0EUYNOB446dal2nPHFCqfXigNCDqealQYNSKoHoKkA5x/Sg BHfK9Kqu3OMVdZeM9qquuDxQK5CWGKjzz3NTPnHtTCOeM0D0E3evUe1APPenhc+1OAANAaEec 9jTt2PxpxHvS7evrQAwtnoKCwyetO/HAoZcc0BoMLfNx0oV+fenDpk1HzmgNCYHvg0ofnGDTA 1SBc9DmgB6MSemBUgYkd6RF468n2p+MKMCmK405IzioyeDgdOasAfjTJCB2GTQBXJ5OQRTS+B 92nOcnk1GxHOMUhhvPapYz8vJ61EvtT1NArllTmn9KgXnGTUq5zTEx2MrTCvFTBMf/AF6UgL9 KZJGw7VVmUOjK4DIwwQRkH1zVogk/SopQuOaks4zXWv01DTrKzjM7Bnljkc/LGu0qC59ixx68 e9W7mCPS9BuA+bl5Bh2fAMzthcn06j6AVvSIM5ORVe5tobmExXEaTRv95WGQfwpcppz3Oe0uW a1to9Lj51FgXknHIYdDLnvngYrXk0+JtMkso8KjRlBnnk9D+fNXBCiEYVRgY4HQU7aB0H/16S VglJN6HMapfxXNpp1vdSKhlm/fhzyBFyw98sF/OtbTommu31CaPbI67IwwwwTOeT6k9u3FOh0 q3S5e6ljR7pyT5hHQeg9Og6VoRRjAotqDmuhm61ILO60/UWwEhk8qU9MJJgZPsGwfzqvoVrZS Q3euX0Fsv2t/MUyRr8kS8L26nlj7sK1dTsU1DT7izmPyTxtGSO2RjP4dfwqpo2htBb2kF/cfa 0s0VYU2BVyAAGYdz+g9KOXUOZctjO1jTlv7KXUZka3SMr5MYVVIjDDcxBBGT+mKNQ0K2k1nTr Nri8mt4kluJEachVGNicKBjJJ/AV1l1bpcwSQyD5JVKkfWuS0q31i8vbh7qF7ZiUiaeRfvJGO NoJ5yxZs9OlD3KjK63HanoFi9zHa24uAzYaQi4k2onpjPU9MVQ0jT7iy1zV202RHjjdIxFdSS YGVBIBBPTpyK7C3sobONkh3Fm5eRzlnPqT61yc8pt5tYtYm23VzeLHGSfuh0B3fQKGP4VL9Co O90RDUrz7Rc3xsgxnxawNDMHBIJ+YZwSM5/Km2OrQ6Fb2zzROYLh3MtxgnaQSBkjjtWnpUCTy rMseLSBRFbAnqBwX/GnaeY7fQjJMMiASAg99rN/h+tFgumM0mSO81zUbuBhLAFiSORTlc7SSR /Kpbu6+wXsscAG+fy9q9t7MVyfw5P0rCg8Py2dnFf6fdfYbth50oPMRzycr2xnFR3sN1faUNS vtsTCaNiIHYBowQAR6dScUa7A4xvozY12S60+GCSxmuLi+mlWNY85V/X5egwOeK0ZIIop9NWN QJvO3Fj94gKdxJ/LNYUmnSf8JLbRw3lwRFA8oM2JNvO0da0oIL6e51CX7RHIYY2t0PlbTuxk4 545wM+1CuDS2ND+15P3DxW3mQzuYoX3cluxI7KcHn2pbrVzaSmCURec7okRyQrlue/oBXNySa jc6f4bVLW2RfPiaN1uCc4U8EY4B5FWbqGae/tbzVI7aSya4MbKMsM7doJz2zmi+guWJ0S3dyu oi0P2d98RlD8jGCByPqevtTzqHkTzpMA23YE8teWZs/KM9/8a5q10z7Preq3Wn3RtFhWOHBG9 Om5hgnPU9iOadcw33n6ZJJMltcSzSfMq713FflBU+wxjOfenfQOWLZ0r6r5V1HbzwuHlGUEfz n8RjipYrmOcnyz8y/eVgVYZ9q4yeXUx4ugWWMS+RaiZjaYO4bmUZVsH14BqG58RynUPt0SSQ2 jSRwAyRf64ZO7rzkDJ46Y5pc4OlfY667u3F19nt40eTbvYu20KOg7E806G7jaNnc+UUba4Y/d NY0Wp2cWv30dxdQRO6RhA7BdwwelV7OVjrup73WS2gVZwwAxuYdOOOAKfML2d1Y6M3MCQ+c0q CP+/nin29zFcpvgkV0zjIPeuOi1FrTVFXUos2cSK6uuT5bPzucf1HArR027kutSv5IyI7TcwE iDO7ZxnP4nmjmE6Z0/BI9KViBwKx7Dzr2x+0zTmJCpKCM44H8R/wAOwo0eZ5rSylaQuZkYt9B 0OKq5PKajEZ5pSQy9ayJ7iWWG8ukk2RQMyqigHft4Yn+mKel1KskLyGMRyqWwB93AB60XBwNM kYI4qM4z2rGnvpntzcxnYOsUYAOVB5LfWnajNIzxW7YCysvKE9M9DSuNRNMtznNWI2z6YrnJr o2F3NAsu6KKNZdsmSwBOMeuK0EvWSDz5INsI5J3jcB6kUKQOBtRtjqe9S59OlZiajaLE0jSqF Hr1OemB3qx9sCbGliljjbjc2OD2yO1O5DjYtM/OFIzULNnIqIXILymRCixoHLHrjBPIqmLmXb DPNsSKY4CYwVXGQc+tNuw1G5cYjqT1pjtgHkU2CUTNIrIQy44J7HpVa51C2hikmcObdPvSKmV H+P4UNisWQ4B6ipk/Q1XUxTRJLCQyMAQw5BFTJwaEFiynQVIp9u9QKxyBxT8kD1Bpk3LAbntR ncKjXANPGf0phYc3t0NRN93qeamP3expjDjgHg0hlQnk/ypjLnpUko+c1FyODSAaMYwaYOaeW w49D1o/i9jQMTHripUBPPNN5xxzUiZwM4NAhVX5vmqdBzhaYox0qVTjqRj0pgxc84PFBCgenv 60vUn1oOOnFAFeUfXmuc1bw5Z3+qxX7tMsqLsYK2FcdOfw4+hrpJMZPSoH+9ikNO2xVWNYwio oCqMAAcCub1cyJDNpkEbSSz3AIA6BCQxye3QiurI59aiwA2e59utJoalYxzp9xeSA6i8fkr0t 4s7CfVieW+nSr15ZrdWU1u+NkiFD7e9W9uTTyM9KYOTZxWhX002o3AI338cKWxT+66k7mPt0P 412um2wtbRYh82B8zEfeJ6n8Tmq9lp8Nvc3M6LiW4YNI3ckAAfhgVqxjjGOKSVhznc5GHTLx9 YgsGjkj0+zeWZZxgbg4wqj3GWGa39WtVk0i4gwAgjIA9McitPoOmaSUZU9KLE87bOK8J79R08 STKQjTPJLu/icHAH0GP5Vp+I4XfTGliGZbd1nUeu3kj8q2iuOmPwqKRA6srDKkEEHvSt0K5tb nG6MJdcutUvYC8dpOfJWToWjReFX0yxYk1SNhNC1s1tJ9ptrWxkkME78Rl8odh/Pr613mm2sN lbRwWsSxwIMKijgVg6RpF1/a2pLcLssjKu1j1lUchR7ZJzScTRTvcg0xFluzFdWzASW0bMk6h sMvGe+RisfR9JtdUkmuVtVjhmnZt65T5B8qqoGPQkn3rtdT01Z/OYPJG8sflb0PIGckii1tUt 7aO3iGEjUKvsKfKT7TQwrloNOu76SQZje1DkH5gdny45+orKisbq00OSOwlb5oma4jHKKxGTt 9Dz06V0WtaONQNtufascgZx/eXOSv5gVoCBQm3GF6YxxS5QVSxyb2c955OnQ6zefYzbia4LhC FQ8KoIUHJIP4A1r2/hxLi0jSe5v/JjGIQJihHOc4XGOnSqvhTS5kkuHmz9mabKLIPnKrwgPsO v412aqOQeKcY6ahOo72R53dW11ZaLqqwarfeZHM0UcR2EOzEBeq55yKLnStVtrDybLUQba0iK fvIRuzjkDHb696v62k6+I4ooY2eN5Uuc9tyoQM/8CI/KujS2RLYRc7Qu05/WkolOdjhdVTVNu mpvguo5pECiL907Ko3kHJx265q5qRh1GRZb62ngSB1TMj468typ9KeEuP7U060SImS0SY78fK AflVufx96ua9ps9zpy2FrGCjsN8meevPHv3NLlDm2uc1ZaRBqWn3N4Y5DGxL/O7HIH3VHPQDn 6mtDUre9eIabZ3pkgkQSSrOf9XH6Bxzljxg10/wBkSLSntokGBEUAH0rN8O6fMtuLnUIytxId 5Q9j0GfoMAfjTcR85BqNr5lrBfMrWlwskSIRztTIGCOhHfFT6y+olY7Ke7smR8ytIyNGVVCOp B4JOBxVrxJEZNDutrrGVUSBmOACpB5/Kqeg2suqXh1e/R0VsG3hJ+6o5Un35JA96LdAU9OZmh qz6kdGnlhsYz5gzJE0vzhcDp2yPc1Th1m01PWLeLz1t4oYg0kU42Mzt0Az2AGffNdPcIZLGZe pZGAz7qa5LRQbwX0Hlo88sqKzSIGEaLGgBwep64HvTa1JjJNal7Wbry4J1t8PNdSpCi7sFx0Y g/Tv71Q8Q6rFLoUtvChiuA8cT20o2soLAfiOnTip9d02LRtNtriygAitZ45JcnLbAecfTOcD3 qt4uhj1NdMiiw6TzL5cyEbl6NkHt8oNTLqONtGaN082jWmIAJImIREbqjN0/DNPnFxp9ubiW8 klMYLyKwGwjvgY49jWDrjanHJDbrKl1FA6yNK6YZCThd+D83fpVvUP7Vn1CC1K2Nwq4uJFJdB tU4VTweCef+A03uPlVtzeS/eJY5LjASVSyqAdw9B7k5FT/arhCktwsccLNtIzllz0yeh57VhM 2o3ctpcFbe2cuyKCTKvQnd264NO1C7vZ8afc2BkdgHd7R9+I84BAbBBJ479z2ppsz5bnSwXOX VWjZAyl1LdwP5VYR9655wehPesK0vre9WdYiySRoIvKkG2Rc+o7/XnOK3AAqgEEkDFWiJKxYb k8cCkbJXj+dN464NNyM4OeaLCIZT82PbvULbgvTr3qw6gDv6VA0fagegwDJ+lPUA4BpgVT17d qkVFx60gFA5xip0AHBpgAPb9acIwRz0oFYfw33e1KTnkc0Io5GKUIq/8A66YCKex/Olc4OD+d NYDdnp+NNO0jPekAjkbqhdhT5FBAxUTRqDg5+uaB6DSck0DG4UoQYpdgB7/nQAg5Y08DIpRGu epFO2DHGadgHRgZ6AVKOAAMCmLHjPWpAuRmixNhQRnjFNkOetOKjHNM2gjnjFFh2I25zxUbA8 5OM9qldfxphRaABRwPWpY+DzUKoM9P/rVKqjvSASdN4OOOKpMjK/Jq9jK96idB36UWArD604A Y44NBUbuOlJx6UDJIhtbI/nVpTkHNVVUHpU6oPxPegRHNAjSrKFy6ghT6A9cUMv0FT7ABjr70 joD2+tFgKLquTzzTQvYGrDRBT0puwelAyIoDQqgDHAFTFMdqQoO64oAhdEdSrYZTwVPQ1NFGu V6cen+FCxgdhzVlFX0H5UxCTDMLKhAJXAz0rK0DTH06C4EsiyyzTNIzKuBjjA/ACtmRVx2pu0 gdulK3Ud9LFeVQyFWwVPBz3rhLWf7J4kubRImk+yM32eEH+KTHT0AAPPbNegSJntWcun28V5N cxwotxMAHkxywAwKHG5UZ2uZN9ZFNEvfNZWuJFMrsOBkdMewxgUzQoH1F7i/uS6pcEBIz8pMa jgn2JJOPet6S2jkxvQMAc4PSpoIljQKigKo4UcUW1By0Kl5FhYWVf9XIp7YA6fyqppGLuW6nX lZ5Su8f880+VcH35NbTRhwQwByOQeadDGsaBVUKqjAAGAB9KdiOayI3tIQocRorIu1SF5Uen6 VZUce1OxkHPehQcAkU0LcZnP40A89KSgdaBj8jrjNQuMjmpe1RN/ShgRN04po4wBwM040wdvr SGTAkEY6elSI2T0qP0p0f3qALCjn60uAD3NC9qD978aaExpPPHFRE/MRzx71L/FUJ/wBYaAEf pUMhywHr1qZ+lQP98UDHDpSqaQfdNKvakBIcLzT1wQRjgUx+lPj6t9KYiRBkn2pScA4xSR9TQ 3Q/SgB7E/pTD9005/6U3+E0ARt1xTQecU5vvCm/xUMA6Y4pVNIf4frQOlIBxYheKicncBUjfc qN/vimBAaRetKaROv40hjkODVlW/Wqo61YX+tAEopD96lWkP3j9KaENbliKaacfvmmmgQ1+lI ORQ/Q0LSGPjFTrwKhjqYfd/GmA3JyBntTwc/jUff8KevQfSgBrDLVG4AOMCpf4vx/pUUn3qAE AxzTxjbnApvb8KcPuUgF2gCngDjimmnDoKZBIOh9qTsTSj7rUn8NBSD/2Q== ---MOQ1109720502ee4e892280cb5c59e0665b145f4a45dd-- From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 1 21:33:29 2005 for ; Tue, 1 Mar 2005 21:33:28 -0500 (EST) for ; Tue, 1 Mar 2005 21:33:28 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 1 Mar 2005 21:32:23 -0500 02:32:17 -0000 j222WHtH008082 for ; Tue, 1 Mar 2005 21:32:17 -0500 (EST) 21:32:17 -0500 (EST) Date: Tue, 1 Mar 2005 21:32:17 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Some Practice Translations To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello again, Here are some practice English -> FOL translations. There are many representational choices you could make for these sentences, but I would like you to pick from the following menu of syntax/semantics: Constants --------- [[logic]] = logic [[albert]] = me :) Predicates ---------- [[Talented(x)]] = [[x]] is talented [[Student(x)]] = [[x]] is a student [[Likes(x,y)]] = [[x]] likes [[y]] [[Man(x)]] = [[x]] is a man [[Woman(x)]] = [[x]] is a woman [[RealNumber(x)]] = [[x]] is a real number [[Between(x,y,z)]] = [[y]] is between [[x]] and [[z]] [[Fly(x)]] = [[x]] is a fly [[Flies(x)]] = [[x]] can fly [[FlyOn(x,y)]] = [[x]] flies on [[y]] [[Plane(x)]] = [[x]] is a plane [[TA(x)]] = [[x]] is a teacher's assistant Represent the following: 1. Every talented student likes logic 2. Some students who don't like logic, don't like anything. 3. No flies like logic. 4. All talented students like some untalented TA. 5. Albert likes all students who like logic 6. Between any two real numbers there is a real number. 7. All flies can fly. 8. No man or woman can fly without some plane. Answers below . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Answers: Since my keyboard doesn't have the rotated A or rotated E, I will use A and E to represent the quantifiers. I only add the "correct" number of parens when there is ambiguity. 1. Every talented student likes logic Ax((Talented(x) ^ Student(x)) > Likes(x,logic)) 2. Some students who don't like logic, don't like anything. Ex(Student(x) ^ -Likes(x,logic) ^ Ay(-Likes(x,y)) 3. No flies like logic. -Ex(Fly(x) ^ Likes(x,logic)) 4. All talented students like some untalented TA. Ax((Talented(x) ^ Student(x)) > Ey(TA(y) ^ -Talented(y))) 5. Albert likes all students who like logic Ax((Student(x) ^ Likes(x,logic)) > Likes(albert,x)) 6. Between any two real numbers there is a real number AxAy((RealNumber(x) ^ RealNumber(y) ^ -(x=y)) > Ez(Between(x,z,y) ^ -(z=x) ^ -(z=y) ^ RealNumber(z))) 7. All flies can fly. Ax(Fly(x) > Flies(x)) 8. No man or woman can fly without some plane. Ax(((Man(x) v Woman(x)) ^ -(Man(x) ^ Woman(x)) ^ Flies(x)) > Ey(Plane(y) ^ FlyOn(x,y))) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 2 10:19:31 2005 for ; Wed, 2 Mar 2005 10:19:31 -0500 (EST) for ; Wed, 2 Mar 2005 10:19:31 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 2 Mar 2005 10:16:26 -0500 15:16:25 -0000 for ; Wed, 2 Mar 2005 10:16:25 -0500 (EST) -0500 (EST) Date: Wed, 2 Mar 2005 10:16:24 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: A correction for the Practice Translations To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Dr. Rapaport caught a bug in my answers to the Practice Translations The answer given to #4 4. All talented students like some untalented TA. was Ax((Talented(x) ^ Student(x)) > Ey(TA(y) ^ -Talented(y))) it should instead be Ax((Talented(x) ^ Student(x)) > Ey(TA(y) ^ -Talented(y) ^ Likes(x,y))) It worries me that I forgot to make you talented students like me :) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 2 16:37:39 2005 for ; Wed, 2 Mar 2005 16:37:39 -0500 (EST) for ; Wed, 2 Mar 2005 16:37:39 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 2 Mar 2005 16:37:05 -0500 21:37:05 -0000 j22Lb5x3009709 for ; Wed, 2 Mar 2005 16:37:05 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 2 Mar 2005 16:37:05 -0500 (EST) Date: Wed, 2 Mar 2005 16:37:05 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: ANSWERS TO 4 TRANSLATION PROBLEMS FROM TODAY'S LECTURE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject:CSE4/563: ANSWERS TO 4 TRANSLATION PROBLEMS FROM TODAY'S LECTURE ------------------------------------------------------------------------ ======================================================================== Here are my representations of the 4 sentences I put on the board at the end of today's lecture, together with my explanations. ======================================================================== 1. If Abe is taller than Bea, then no one is taller than Abe. In any translation/representation into FOL (or any other KR language), you need first to decide what your predicates are. This really should be done by first having an ontology (a theory of the kinds of things in your domain). On the exam, I will provide this for you, but, in real life, you have to do it on your own, or rely on someone else's ontology. For this example, we need 2 individuals, Abe and Bea, and a 2-place relation of being taller than. Thus, our language will need 2 constants, a,b, and a 2-place predicate: Taller. [[a]] = Abe [[b]] = Bea [[Taller(x,y)]] = [[x]] is taller than [[y]] The only other potentially tricky part is how to represent "no one is taller than Abe". If no one is taller than Abe, then there isn't anyone who is taller than him, so it's not the case that there is someone who is taller than him: -Ey.T(y,a) Final translation: Taller(a,b) > -Ey.T(y,a) ------------------------------------------------------------------------ 2. All owners of books read some fiction. Here, we need to be able to talk about an owner of a book, so we need to be able to talk about owning something, so we need a relation of owning. And we need the property of being a book. We also need the relation of reading something. More problematically, we need to decide what it means to "be fiction". Here, I will simply introduce a predicate that means "is fiction" without having to analyze what "being fiction" means exactly. [[Owns(x,y)]] = [[x]] owns [[y]] [[Book(x)]] = [[x]] is a book [[Reads(x,y)]] = [[x]] reads [[y]] [[Fiction(x)]] = [[x]] is fiction. Ax[ Ey[Book(y) ^ Owns(x,y)] > Ez[Fiction(z) ^ Reads(x,z)] ] I.e., all owners of books are readers of fiction. Notice that the antecedent says "x owns a book" (i.e., there is a book that x owns; i.e., x is an owner of a book) and the consequent says "x reads fiction" (i.e., there is some fiction that x reads; i.e., x is a reader of fiction). ------------------------------------------------------------------------ 3. Every doctor who treats himself has a fool for a client. Here, we need(*) to represent being a doctor, treating someone, being a fool, being a client of someone: [[Doctor(x)]] = [[x]] is a doctor [[Treats(x,y)]] = [[x]] treats [[y]] [[Fool(x]] = [[x]] is a fool [[Client(x,y)]] = [[x]] is a client of [[y]] Ax[ (Doctor(x) ^ Treats(x,x)) > Ey[Fool(y) ^ Client(y,x)] ] I.e., all doctors who treat themselves are such that they have a fool for a client. Note that the antecedent says that "x is a doctor who treats him/herself" (i.e., x is a doctor and x treats x) and the consequent says that "x is such that a fool is his/her client" (i.e., there is a fool who is a client of x). By the way, it does *not* logically follow from this that the doctor is the fool (although that's surely the pragmatic implication)! ------------------------------------------------------------------------ 4. If anyone gives me a stupid gift on my birthday, I won't send him a thank-you card. [[Gives(x,y,z,w)]] = x gives y to z on w Note that this is, perhaps, an unusual use of "give", which is normally a 3-place predicate (x gives y to z). Giving "on" an occasion is not the usual meaning, though it's not uncommon, either, but adding this 4th argument place to "Gives" makes it easier to represent this sentence! [[Gift(x)]] = x is a stupid gift Why not have a "gift" predicate *and* an "is stupid" predicate? You could; there doesn't seem to be any special reason to do so for this one, isolated sentence. But there's another problem having to do with adjectives: Although a stupid gift might be both stupid and a gift: Gift(x) [in the above sense] == Stupid(x) ^ isGift(x)), in general this isn't the case: A toy gun is not both a toy and a gun (in fact, it is a toy, but not a gun), a small elephant is not both small and an elephant (it is an elephant, but not small, although it is small-for-an-elephant), and an alleged murderer is not both alleged and a murderer (until proven guilty), though an alleged murderer is "alleged" (whatever that means exactly) but not necessarily a murderer. [[Birthday(x)]] = x is my birthday Note that including "me" in this predicate makes it easier to represent; again, in the absence of further context, there seems no reason to go into more detail in the representation. If you did want to represent that some day is a birthday and that it is *my* birthday, you'd need to decide how to represent possession. Perhaps: Birthday(x) [in the above sense] == is_a_birthday(x) ^ possessor(x,m) (where [[m]] is given below). [[m]] = I (or me, depending on context) Since we do need to represent "me", perhaps we should have represented it in the "Birthday" predicate as shown above, but doing so wouldn't really add much to the representation, so I won't. [[Send(x,y,z)]] = x will send y a z (i.e., x will send a z to y) Note the inclusion of the future tense in this predicate. [[TYC(x)]] = x is a thank-you card At least we don't have to come up with a theory and analysis of what a "thank-you card" is! Our sentence becomes: Ax[ EzEy[Gift(y) ^ Birthday(z) ^ Gives(x,y,m,z)] > -Eu[TYC(u) ^ Send(m,x,u)] ] Here, the antecedent says that "x is a giver of a stupid gift on my birthday" (i.e., there is a stupid gift and there is my birthday and x gives that stupid gift to me on my birthday). And the consequent says that "x is such that there's no thank-you card that I will send to him/her" (i.e., there's no thank-you card that I will send to x; i.e., it's not the case that there is a thank-you card that is such that I will send it to x). So, the sentence says: all givers of stupid gifts on my birthday are such that there's no thank-you card that I will send them. ======================================================================== (*) Actually, "need" to represent is too strong. We *can* represent the sentence by representing these properties and relations. There may also be other ways to represent the sentence with a different ontology. In the absence of a particular ontology or any other context (such as other sentences that might be part of a discourse or argument), there's no single answer to what "needs" to be represented. ======================================================================== From stock@buffalo.edu Wed Mar 2 22:40:47 2005 Wed, 2 Mar 2005 22:40:46 -0500 (EST) In-Reply-To: <200503022137.j22Lb5kI009708@wasat.cse.buffalo.edu> Cc: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU From: stock@buffalo.edu Subject: Re: CSE 4/563: ANSWERS TO 4 TRANSLATION PROBLEMS FROM TODAY'S LECTURE Date: Wed, 2 Mar 2005 22:40:45 -0500 To: "William J. Rapaport" On Mar 2, 2005, at 4:37 PM, William J. Rapaport wrote: > Ax[ Ey[Book(y) ^ Owns(x,y)] > Ez[Fiction(z) ^ Reads(x,z)] ] In this one, I added a Book(z) to the Fiction(z) ^ Reads(x,z) clause. Would that be "reading" too much into the problem? > Ax[ (Doctor(x) ^ Treats(x,x)) > Ey[Fool(y) ^ Client(y,x)] ] In this case, I didn't use a predicate for Client(), instead reusing Treats(). Again, is that too loose for a literal translation, given that we're trying to replicate the exact text and not the exact meaning? Thanks, -Matt From rapaport@cse.Buffalo.EDU Thu Mar 3 09:09:40 2005 for ; Thu, 3 Mar 2005 09:09:40 -0500 (EST) for ; Thu, 3 Mar 2005 09:09:40 -0500 (EST) Thu, 3 Mar 2005 09:09:39 -0500 (EST) Thu, 3 Mar 2005 09:09:39 -0500 (EST) Date: Thu, 3 Mar 2005 09:09:39 -0500 (EST) From: "William J. Rapaport" To: rapaport@cse.Buffalo.EDU, stock@buffalo.edu Subject: Re: CSE 4/563: ANSWERS TO 4 TRANSLATION PROBLEMS FROM TODAY'S LECTURE Cc: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | From stock@buffalo.edu Wed Mar 2 22:40:47 2005 | Subject: Re: CSE 4/563: ANSWERS TO 4 TRANSLATION PROBLEMS FROM TODAY'S LECTURE | | > Ax[ Ey[Book(y) ^ Owns(x,y)] > Ez[Fiction(z) ^ Reads(x,z)] ] | | In this one, I added a Book(z) to the Fiction(z) ^ Reads(x,z) clause. | Would that be "reading" too much into the problem? I think so (even ignoring the pun :-). After all, if every book owner reads some fiction, it's possible that the fiction that the book owner reads isn't in a book, but in, say, a magazine or on the web. | > Ax[ (Doctor(x) ^ Treats(x,x)) > Ey[Fool(y) ^ Client(y,x)] ] | | In this case, I didn't use a predicate for Client(), instead reusing | Treats(). Again, is that too | loose for a literal translation, given that we're trying to replicate | the exact text and not the exact meaning? Yes. Although I agree that there's a clear implication that the client is someone who is treated by the doctor, in the absence of any further context to clarify this, it's possible that the doctor is also a lawyer and the the foolish client has sought legal, not medical, advice. The general issue here is what to do with isolated sentences that are not in a surrounding context. In general, I think the rule of thumb should be to represent such sentences as literally and as "narrowly" as possible, not "reading" too much extra information into them. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 3 09:11:59 2005 for ; Thu, 3 Mar 2005 09:11:59 -0500 (EST) for ; Thu, 3 Mar 2005 09:11:58 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Mar 2005 09:11:00 -0500 14:11:00 -0000 j23EAux3013209 for ; Thu, 3 Mar 2005 09:10:56 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Mar 2005 09:10:56 -0500 (EST) Date: Thu, 3 Mar 2005 09:10:56 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: ONTOLOGY CONFERENCE @ UB To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: ONTOLOGY CONFERENCE @ UB ------------------------------------------------------------------------ There will be an ontology conference at UB that is highly relevant to KRR: | National Center for Ontological Research | NCGIA and Department of Philosophy | University at Buffalo | | present | | MAPPING THE HUMAN BODY | | April 16-17, 2005 | | Presentations will include: | | Olivier Dameron (Stanford University/University of Rennes): | Spatial Relations in the Human Brain | | Olivier Bodenreider (Medical Ontology Research, National Library of | Medicine, Bethesda, MD): | The Body Region Connection Calculus | | Maureen Donnelly (IFOMIS, Saarbrücken): Representing and Reasoning in | Biomedical Ontologies | | David Mark (University at Buffalo): | Hills, Noses, Cliffs, Heads: Earth Parts and Body Parts in Language and | Thought | | Anand Kumar (IFOMIS, Saarbrücken): Ontology of Pathological Structures | | Barry Smith (University at Buffalo): Wounded Knees | | A full program is available at: | http://ontology.buffalo.edu/anatomy_GIS/ | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 3 19:43:48 2005 for ; Thu, 3 Mar 2005 19:43:48 -0500 (EST) for ; Thu, 3 Mar 2005 19:43:48 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Mar 2005 19:40:02 -0500 00:40:02 -0000 for ; Thu, 3 Mar 2005 19:40:01 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Mar 2005 19:40:01 -0500 (EST) Date: Thu, 3 Mar 2005 19:40:01 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: Re: Some Practice Translations To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | I have a (n admittedly last-minute question) on one of the answers given | below: | | Quoting Albert Goldfain : | | | > 8. No man or woman can fly without some plane. | > Ax(((Man(x) v Woman(x)) ^ -(Man(x) ^ Woman(x)) ^ Flies(x)) > | > Ey(Plane(y) ^ FlyOn(x,y))) | | | I'm not very happy with any of my solutions to this problem, but here is | the answer that seems to work the best for me: | | Ax(((Man(x) v Woman(x)) ^ Fly(x)) -> EyPlane(y)) | | (oof, my email client doesn't do paren matching...I hope there are a | correct amount of them.) | | In keeping with the guidelines delimited in previous emails about books, | fiction, doctors, clients, and lawyers, I don't understand why the | "FlyOn" predicate is necessary when that's not explicitly what's stated | in the sentence...that's just what we know about the world. I readily | admit that my formula does not echo the explicit form of the example | either, but it seems logically equivalent to its English counterpart at | least, and it's not confounded by what we know about the world. | | More generally, I guess I can't derive, given previous examples what | level of correspondence we should keep between our representations in | FOPC and their English sentences, as we were given a 1-place predicate | for things like "plays with a gorilla" for the sentence "No kitten | without a tail will play with a gorilla", or "loves fish". I would have | put a two-place predicate in these sentences, and then quantified | gorillas and fish as well. | | Any advice on how to deal with these translations, both in terms of the | exam tomorrow and for general purposes later on? For the exam, don't worry, since I'll provide the semantics :-) For general purposes, I think you have a good argument. The point is that without further context, there's really no single best "correct" translation, though there *are* incorrect ones, and there's no way to decide which of several equally "correct" ones are better than others. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 3 19:45:31 2005 for ; Thu, 3 Mar 2005 19:45:31 -0500 (EST) for ; Thu, 3 Mar 2005 19:45:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Mar 2005 19:44:45 -0500 00:44:18 -0000 for ; Thu, 3 Mar 2005 19:44:07 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Mar 2005 19:44:07 -0500 (EST) Date: Thu, 3 Mar 2005 19:44:07 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: Translating from Propositional Logic to English To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU I wrote: | > ------------------------------------------------------------------------ | > Subject: CSE 4/563: Translating from Propositional Logic to English | > ------------------------------------------------------------------------ | > | > 1. | > > | > Let [[snow]] = It's snowing. | > and [[sunny]] = It's sunny. | > and [[buffalo]] = We're in Buffalo. | > | > > 3. | > This is not directly related to any of the above, but here are two | > examples of grammatically correct and meaningful, but hard-to-understand, | > sentences of English. Can you figure out what they mean? | > | > Dogs dogs dog dog dogs. | > and | > Buffalo buffalo buffalo buffalo buffalo. | > | And a student replied: | Are we to be interpreting "Buffalo" just above as [[buffalo]] denoted in | (1)? No; sorry if I misled you. Here's some background: The "dogs" sentence is a perfectly grammatical sentence of English needing no further punctuation. Can you figure out its meaning? The "buffalo" sentence has the same grammatical structure. Both sentences are perfectly grammatical and meaningful (albeit odd :-). | Otherwise I think I might need the statement said out loud, to be | able to pick up on intonation patterns. Or | perhaps I'm reading too much into the whole thing. If I were to say it out loud, I'd probably stress the 2nd and last words in each sentence, with the heavier stress on the 2nd. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 3 20:29:40 2005 for ; Thu, 3 Mar 2005 20:29:39 -0500 (EST) for ; Thu, 3 Mar 2005 20:29:39 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Mar 2005 20:28:45 -0500 01:28:45 -0000 for ; Thu, 3 Mar 2005 20:28:44 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Mar 2005 20:28:44 -0500 (EST) Date: Thu, 3 Mar 2005 20:28:44 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: CSE 4/563: Translating from Propositional Logic to English To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A student writes: | > | > Buffalo buffalo buffalo buffalo buffalo. | > | > | | Actually, a friend of mine pointed me to an interesting mini-article about | this very topic just a few weeks ago, and I've been bugging my friends with | sentences consisting of "buffalo" x N ever since. Let me find the URL... Ah. | http://itre.cis.upenn.edu/~myl/languagelog/archives/001817.html . It's got | an expansion of the first six sentences consisting of "buffalo" repetitions, | and it's pretty entertaining. Neat; I hadn't seen this site, but its author (Geoffrey Pullum) is a well-known linguist (indeed, a computational linguist even). There's an interesting history to this sentence. It first appeared in print as far as I know, in a book by another computational linguist, Robert Berwick (_Computational Complexity and Natural Language_ is the title, I think). He has told me that he first heard it growing up in NYC in the 1950s. It has the same grammatical structure as "Police police police police police". What I find interesting about this sentence is that at least two people, widely separated in time and space, have independently generated this sentence: When I was in grad school in Indiana in the 1970s, I devised it as a "purer" version of "Dogs dogs dog dog dogs" (which one of my professors, John Tienson, had used in a philosophy of language course)--purer in the sense that there is no plural -s marker on the noun. It is also cited, though with a different grammatical structure, in Steven Pinker's _ The Language Instinct_. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 3 20:53:13 2005 for ; Thu, 3 Mar 2005 20:53:13 -0500 (EST) for ; Thu, 3 Mar 2005 20:53:12 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 3 Mar 2005 20:52:43 -0500 01:52:42 -0000 for ; Thu, 3 Mar 2005 20:52:42 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 3 Mar 2005 20:52:42 -0500 (EST) Date: Thu, 3 Mar 2005 20:52:42 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: More on "buffalo" (and why aren't you studying?) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: More on "buffalo" (and why aren't you studying?) ------------------------------------------------------------------------ Buffalo buffalo buffalo buffalo buffalo. [[is-a-buffalo(x)]] = x is a buffalo [[buffalos(x,y)]] = x buffalos y Ax[(is-a-buffalo(x) ^ Ey[is-a-buffalo(y) ^ buffalos(y,x)]) > Ez[is-a-buffalo(z) ^ buffalos(x,z)]] From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 7 09:36:57 2005 for ; Mon, 7 Mar 2005 09:36:57 -0500 (EST) for ; Mon, 7 Mar 2005 09:36:57 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 7 Mar 2005 09:36:43 -0500 14:36:43 -0000 j27Eagx3001064 for ; Mon, 7 Mar 2005 09:36:42 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 7 Mar 2005 09:36:42 -0500 (EST) Date: Mon, 7 Mar 2005 09:36:42 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: F.O.L. WEBSITE UPDATED To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: F.O.L. WEBSITE UPDATED ------------------------------------------------------------------------ I have added some items to the FOL website: http://www.cse.buffalo.edu/~rapaport/563S05/fol.html 1. A "crash course in the theory of logic" by a well-known philosopher and computational linguist, Rich Thomason 2. Two papers on the relationship between FOL and computer science/software engineering, one by Joseph Halpern et al. and one by Peter Henderson (on education). And I have added a webpage on Formal Systems to the Syntax/Semantics website; go there directly at: http://www.cse.buffalo.edu/~rapaport/563S05/formalsystems.html (This page explains what a "formal system"--i.e., a syntactic domain--is, discusses the relationships between syntax and semantics, and provides a webliography of further information.) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 8 21:33:07 2005 for ; Tue, 8 Mar 2005 21:33:07 -0500 (EST) for ; Tue, 8 Mar 2005 21:33:07 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 8 Mar 2005 21:32:12 -0500 02:32:12 -0000 for ; Tue, 8 Mar 2005 21:32:12 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 8 Mar 2005 21:32:12 -0500 (EST) Date: Tue, 8 Mar 2005 21:32:12 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MIDSEMESTER COURSE EVALUATION SUMMARY & COMMENTS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MIDSEMESTER COURSE EVALUATION SUMMARY & COMMENTS ------------------------------------------------------------------------ Thanks for your comments on the midsemester course evaluation. There were two parts: one for the recitation and the TA, one for the lecture and me. I'll review the recitation comments with Albert, and he can address any concerns directly. Here, let me try to summarize and respond to your comments on the course as a whole. In general and on the whole, you seem happy with the way things are going, with a few significant exceptions (with which I more or less agree). A few of you noted some areas where things might have been done differently or areas that I need to take into consideration. ======================================================================== Need for improvement: ------------------------------------------------------------------------ 1. About 20% of you said that you wanted more time for the exam (or to have had the course on a T/Th instead of MWF; unfortunately, I have no say on when our courses are offered). I'll admit that perhaps it was a bit long, but I'll refrain from saying more about that until we finish grading it. You'll certainly have more time for the final exam. 2. And the same number of you think that we're going too slowly. I also feel that I could have done a few things more efficiently or quickly. Note, however, that something that a lot of you *liked*, namely, my clear explanations of topics, might have disappeared had we gone through things more quickly. An instructor has a trade-off to consider: Cover lots of material quickly and shallowly, or less material more slowly and in depth. When there's no sequel being taught by someone else such that this course must end where that one will begin, then the instructor has to cover the required topics. But when there's no sequel, or the sequel's topics are determined by the first course's topics, then one can go more leisurely. I tend to prefer leisure and depth, which, ideally, can give the student the ability to learn more material at a later date on their own. Having said all that, I'll also say that the pace will pick up very soon. Fair warning :-) 3. About 15% of you would like to see more examples in class. OK; I'll try to give more. If you'd like to see an example of something that I don't give an example for, ask for one! 4. About 10% of you would like to cover more and different kinds of logics or less FOL. Here, the problem is that FOL is the basis for all the others, and you need a solid grounding in that first. There's no other course we offer that gives you that, I think. For undergrads, there's a unit on logic in 191, but that's pretty superficial. For grads, I don't think there's anything (unless, of course, you've had a logic course as an undergrad--and I mean a full course in logic or a unit in an undergrad AI course, not a unit in a freshman-level discrete math course). In any case, we *will* be covering at least one more system of logic, that of SNePS. If we have some time, I'll try to discuss some other logics, such as modal or epistemic logics (but note that not even our text covers those!). As for other KRR systems, they'll be covered in 663, or you can read about them in our text on your own! 5. And about 10% of you are concerned about the project (but see #6 in the "things you like" section, below!), either feeling that I assigned it too early or that it should count more. I'm going to ascribe this to general nervousness about it, but I think you'll see that it will be pretty straightforward once we really get started covering the material, probably next week. 6. About 7% of you think I assign too much recommended reading. Wait: If it's only recommended, then it's not really assigned, is it? You need to pick and choose what seems interesting to you to read, keeping in mind that you can always read some of the recommended items later on. (But see #5, below :-) 7. And about 7% of you don't like the text. I'm going to guess that that's because the text doesn't go into as much detail on some topics as I do, so that we're really only covering a small portion of the text. I agree that that's a problem, and I have looked and am looking (so far, unsuccessfully) for other or supplementary texts. In any case, this is The Classic Text in KRR, and a good volume to have in your professional library to serve as a reference later on in your career. 8. One or two of you think that the HWs are too long or that there are too many of them. (But see # 4, below :-) One or two of you think that I use too many symbols. Sorry; that's unavoidable in a formal course such as this. And one or two of you think that I should give clearer reading assignments. Fair enough; I'll try to do better. ======================================================================== Things you like: ------------------------------------------------------------------------ 1. About 40% of you like me and/or my style of teaching or the course organization or the detailed explanations of concepts. Thanks! 2. About 25% of you like the Listserv and/or the website. Thanks! 3. About 20% of you like my logical/philosophical/interdisciplinary approach to KRR. Thanks! 4. About 15% of you *like* the frequency and length of the HWs (see #8, above :-) 5. The same amount like the readings; see #6, above :-). 6. One or two of you like the fact that I've given you a lot of time for the project; see #5, in the "need for improvement" section, above :-) 7. One or two of you like my grading scheme; thanks. 8. And one or two of you like the challenging nature of the course. Thanks! ======================================================================== If you have any further comments, please share them with the class via the Listserv, with me via private email, or anonymously by putting a note in my mailbox in Bell 211. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 11:35:15 2005 for ; Wed, 9 Mar 2005 11:35:14 -0500 (EST) for ; Wed, 9 Mar 2005 11:35:14 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 11:34:46 -0500 16:33:53 -0000 j29GXqx3014086 for ; Wed, 9 Mar 2005 11:33:52 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 9 Mar 2005 11:33:52 -0500 (EST) Date: Wed, 9 Mar 2005 11:33:52 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: COMMENT ON MIDSEMESTER EVAL AND ON PROJECT To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: COMMENT ON MIDSEMESTER EVAL AND ON PROJECT ------------------------------------------------------------------------ This is long, but please read it all for a suggestion on how to get started on the project! A student writes: | First, I'd like to comment that I've never seen a professor try this | much to improve their course. I appreciate that. Thanks :-) | | I think all of the nervousness about the upcoming project is due to a | combination of factors: | | 1) We were given about 5-6 weeks notice, which is intimidating. I | think the purpose is a good one, though - I can't go through a day or | two without thinking to myself "shouldn't I be working on that | project?" :) After this week, when my midterms are all done, I'll | actually start it. But I'll bet you that, for each student who is intimidated by the long lead time, there is a student who would have been intimidated by a shorter overall time for the project. It's impossible to please everyone. | 2) You said we wouldnt get to the material we would need to know until | possibly after spring break. Then, you said this didn't concern you | since we could always read the book and be ahead of the lectures. This | concerns us because we have to worry about learning some unknown piece | of information from the book that we have to master and use in a | seemingly large project. You're quite right, and I apologize for making you nervous (see below), but listen carefully to what you've just said: "learning some unknown piece of information from the book that we have to master and use". In the real world, that'll be the primary way you'll learn new stuff: on your own, from texts (or journal articles). Eventually, you all need to learn how to learn on your own, not just from a teacher; you need to learn how to learn. | I'm sure we're all making too much of it, but | this is how it feels to me and I'm thinking that I'm not alone. I don't think you're making too much about it. I just think you're all understandably a bit nervous about it. | My suggestion would tie into the last item on your improvements list - | ease our worry by specifying what topics/page numbers in the textbook | we should look at for the project. OK--good idea! For the first part, on conjunctive normal form, look at B&L, pp. 50-52, 55-57, 64-67. For unification, look at B&L, pp. 57, 71-72. For resolution, pp. 49-55, 58-61. To summarize, take a look at pp. 49-61, 64-67, 71-72. (Note, by the way, that Ch. 4 goes from pp. 49-75 (not counting the exercises), so this is most of that chapter.) | Perhaps also suggest how to get | started, and we should be fine - the hardest thing on these projects | is always getting started. I'd start by looking at the sections in B&L on CNF & the online algorithm at: http://www.cse.buffalo.edu/~rapaport/563S05/clause-form/clause-form.html Decide which programming language you're going to use, and how best to represent FOL wffs and CNF wffs in that language. Then decide how to implement the CNF algorithm in that language. It will take as input an FOL wff represented in that language, and will output its CNF (as represented in that language). | This course is a good one and I'm sorry to see that your name is | absent from the course directory for the Fall. Thanks and keep up the | great work (both you and Albert)! Thanks again :-) I'm not scheduled to teach next semester, since I'm hoping to be on sabbatical, continuing my writing on the philosophy of artificial intelligence and my research on contextual vocabulary acquisition. Don't forget that the sequel to this course, CSE 663, will be taught next semester by Prof. Shapiro. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 12:08:07 2005 for ; Wed, 9 Mar 2005 12:08:06 -0500 (EST) for ; Wed, 9 Mar 2005 12:08:06 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 12:07:37 -0500 17:07:37 -0000 j29H7btH028890 for ; Wed, 9 Mar 2005 12:07:37 -0500 (EST) 12:07:37 -0500 (EST) Date: Wed, 9 Mar 2005 12:07:36 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Recitation Grades To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, As an indication of your mid-semester progress, Dr. Rapaport has asked me to assign recitation grades. You will find this grade on the front page of the midterms (you will be getting those back today). Here is the algorithm I used to compute this grade: 1. I took the best 3 of Hw3,Hw4,Hw5, and Quiz1, and averaged them (using the university 4 point scale for letter-number conversions...i.e. A = 4.0, A- = 3.7, B+=3.3...) 2. If you received two or more Check+ grades on Hw1,Hw2, and My Extra Credit Hw(everyone who did this got a Check+), then I raised your grade a half letter (i.e., from a B- to a B, from a B+ to an A- ...). Note: these might not have as strong an impact in the final evaluation for the recitation grade. If you didn't hand in Hw1 and you didn't hand in Hw2 then I reduced your grade a half letter (but no one did this...so I am speaking to no one :) 3. I did not take attendance into consideration for this mid-semester grade b/c most of you have made it to all of the recitations (and those who missed were only absent once or twice). I DEFINITELY WILL take attendance into consideration in the final evaluation for the recitation grade. If there are any questions about this midsemester grade (or any other grade) please come see me during my office hours. Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 14:45:22 2005 for ; Wed, 9 Mar 2005 14:45:21 -0500 (EST) for ; Wed, 9 Mar 2005 14:45:21 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 14:42:43 -0500 19:42:43 -0000 j29JghtH008844 for ; Wed, 9 Mar 2005 14:42:43 -0500 (EST) 14:42:43 -0500 (EST) Date: Wed, 9 Mar 2005 14:42:43 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: A matter of life and death :) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, You may remember that in an earlier recitation (maybe the third or fourth) I discussed some proper representations schemes for the following four sentences: Albert is happy Albert is sad Albert is alive Albert is dead I said that thinking about what possible states you want to represent should guide your choice of propositions. So if your choice is: [[H]] = Albert is happy Then -H should be interpreted as "It is not the case that Albert is happy", but that, knowing that there are many human emotions, it should not be interpreted as "Albert is sad" I then went on to talk about states that were more binary in nature (e.g., life and death). So if your choice is: [[A]] = Albert is alive Then -A can be interpreted as "It is not the case that Albert is alive". I also said that it was reasonable to interpret -A as "Albert is dead" if you didn't want to allow the possibility of a state of being aside from life and death. Now: Putting aside theology, philosophy, and the fact that we were discussing WHAT to represent in PROPOSITIONAL logic at the time :) a few of you might wonder why -Alive(x) could not be interpeted as "[[x]] is dead" on the exam The difference here is that you WERE given two predicates Alive(x) and Dead(x) meaning [[Alive(x)]] = [[x]] is alive [[Dead(x)]] = [[x]] is dead So the designer of the exam ontology (Dr. Rapaport) wanted to leave open the possibility of a third state of life... or, more likely, just the ability to say something like Ex(-Alive(x) ^ -Dead(x)) "There is something that is not alive and not dead" Whatever the truth value of this sentence, the important thing to notice is that it can be said! Why did Dr. Rapaport include this possibility? 1. He is a philosopher of AI...we would like to talk about computational agents without talking about states like life and death :) But this is pure speculation. 2. The sentence "Some bugs are dead. And some are alive" might have been said in response to some cartoon. In cartoon ontologies, I am certain there are states other than life or death (any cartoon enthusiasts out there? :) Anyway, if you are given a syntax/semantics with Dead(x) included (or any other predicate indicating a state of being that you consider "binary") then use that predicate! (If you were not given Dead(x), then -Alive(x) would be your best bet). That having been said...sorry for any confusion I may have caused. I certainly did not discuss the issue at this length in recitation. Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 14:52:07 2005 for ; Wed, 9 Mar 2005 14:52:06 -0500 (EST) for ; Wed, 9 Mar 2005 14:52:06 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 14:50:00 -0500 19:50:00 -0000 j29Jnvx3015151 for ; Wed, 9 Mar 2005 14:49:57 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 14:49:57 -0500 (EST) Date: Wed, 9 Mar 2005 14:49:57 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: A matter of life and death :) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | Date: Wed, 9 Mar 2005 14:42:43 -0500 | From: Albert Goldfain | Subject: A matter of life and death :) | | ... | The difference here is that you WERE given two predicates Alive(x) and | Dead(x) meaning | | [[Alive(x)]] = [[x]] is alive | [[Dead(x)]] = [[x]] is dead | | So the designer of the exam ontology (Dr. Rapaport) wanted to leave open | the possibility of a third state of life... or, more likely, just the | ability to say something like | | Ex(-Alive(x) ^ -Dead(x)) "There is something that is not alive and not | dead" | This last is, I think, the best reason not to automatically interpret "alive" as "not dead" (or vice versa): If you don't allow yourself to have both an "alive" and a "dead" predicate, then you literally cannot *say* that, if something is alive, then it is not dead. The best you could do would be to say that, if something is not dead, then it is not dead, which isn't very informative. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 16:27:19 2005 for ; Wed, 9 Mar 2005 16:27:18 -0500 (EST) for ; Wed, 9 Mar 2005 16:27:18 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 16:20:07 -0500 21:20:07 -0000 j29LJkx3015850 for ; Wed, 9 Mar 2005 16:19:46 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 9 Mar 2005 16:19:46 -0500 (EST) Date: Wed, 9 Mar 2005 16:19:46 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MIDTERM EXAM ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MIDTERM EXAM ANSWERS ------------------------------------------------------------------------ Since I didn't really have enough time to go over the last couple of questions, and since I don't want to take any more lecture time on the midterm exam, I have put the answers (in PDF) with some commentary on the website: Go to: http://www.cse.buffalo.edu/~rapaport/563S05/directory.html or directly to: http://www.cse.buffalo.edu/~rapaport/563S05/midterm-answers.pdf From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 9 16:35:31 2005 for ; Wed, 9 Mar 2005 16:35:31 -0500 (EST) for ; Wed, 9 Mar 2005 16:35:31 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 9 Mar 2005 16:34:52 -0500 21:34:52 -0000 j29LYpx3015945 for ; Wed, 9 Mar 2005 16:34:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 9 Mar 2005 16:34:51 -0500 (EST) Date: Wed, 9 Mar 2005 16:34:51 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: UPDATED SYLLABUS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: UPDATED SYLLABUS ------------------------------------------------------------------------ I have updated the syllabus to reflect what we have done and to be a bit more precise about what you should be reading. Go to the directory of documents for the link. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 10 08:54:41 2005 for ; Thu, 10 Mar 2005 08:54:41 -0500 (EST) for ; Thu, 10 Mar 2005 08:54:41 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 10 Mar 2005 08:54:36 -0500 13:54:36 -0000 j2ADsZx3018435 for ; Thu, 10 Mar 2005 08:54:35 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 10 Mar 2005 08:54:35 -0500 (EST) Date: Thu, 10 Mar 2005 08:54:35 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: REMINDER RE MIDTERM TRANSLATIONS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: REMINDER RE MIDTERM TRANSLATIONS ------------------------------------------------------------------------ A reminder: Please send me your *incorrect* translations of English sentences into FOL wffs from the midterm exam, and I will try to explain to all of you why those translations are wrong. (And if I fail, I'll raise your score :-) Thanks. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 10 19:58:54 2005 for ; Thu, 10 Mar 2005 19:58:53 -0500 (EST) for ; Thu, 10 Mar 2005 19:58:53 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 10 Mar 2005 19:57:41 -0500 00:57:27 -0000 for ; Thu, 10 Mar 2005 19:57:26 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 10 Mar 2005 19:57:26 -0500 (EST) Date: Thu, 10 Mar 2005 19:57:26 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUERY ABOUT MIDTERM EXAM To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: QUERY ABOUT MIDTERM EXAM ------------------------------------------------------------------------ A student writes: | I have a slight complaint with problem 7 on the exam. I whipped through | the first 6 problems in about 35 minutes. At that point I figured I would | be able to get | through the last problem quickly, and then be able to go back and re-check | the rest of my answers. However, I was never able to do this because I | ended up spending the rest of the time on problem 7. Basically what | happened for me on problem 7 is I quickly wrote your first answer | AxAy[(Snowflake(x)^Snowflake(y)) > -(x=y)]. However, I looked at it, and I | thought to myself "what if x is equal to y? that would make this answer | wrong! We | can't make that assumption!" "So", I thought, "I'll add in ^ -(x=y) to the | left | side of the implication" (like you did for your second answer). But then I | thought to myself "but, then this doesn't say anything, it's obvious". So | I went back and forth, and ended up trying to construct something that | wasn't a tautalogy, couldn't come up with anything, and in the end blew | the rest of my test time on that one problem. So after all of that | explanation, my complaint is this: For an exam like this that requires | very precise answers, there should be one correct answer that does not | have any caveats. I wouldn't complain so much, except that I feel a little | cheated considering I spent so much time unsatisfied with the answers that | I came up with (for the same reasons mentioned in the answer key) only to | find out that they were the actual correct | answers with caveats. | | PS - I still did get an A, so it's not that big of a deal :) Thanks for sharing your concern. As you point out, since you got an A, it's "not that much of a big deal". On the other hand, it is, else you wouldn't have emailed me :-) I think the ambivalence you're feeling (it is a big deal, yet it isn't) might be due to a dawning realization that often there aren't clear answers to questions (though wouldn't it be nice if there were--if everything were like "What is 2+2?"?). I try not to give such open-ended questions on *final* exams when there's no chance to go over them and learn from them, but on midterms I think one or two such questions are fair game. Question 7 was like that, and to mitigate the impact, I made them worth far less than the other questions. In fact, a student who *tried* question 7 and only got partial credit (i.e., wrote down something that was syntactically correct even if not semantically correct) could still have gotten an A (as, I think, was the case with you). You also mention that "an exam like this ... requires very precise answers". Well, not necessarily: even in the case of the translations, there's going to be room for some variety in the answers (up to logical equivalence, at least). Since there's nothing I can, or need, to do to raise your grade, let's consider what you might take away from this. One thing is the fact that some questions have no clear answers. Another is to write down your thoughts on the exam, even if you think they don't contain "the" answer. As you see, in this case, you would have been correct! I often advise students that, if they aren't sure how to interpret a question, to make some assumptions, state those assumptions clearly, and then answer the question in terms of those assumptions (that's a sort of conditional answer, with the assumptions playing the role of the antecedent in a kind of application of >Intro!). I hope this helps. If you'd like to discuss this further, stop by, either during office hours or by appointment. -Bill Rapaport From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Mar 11 11:28:40 2005 for ; Fri, 11 Mar 2005 11:28:39 -0500 (EST) for ; Fri, 11 Mar 2005 11:28:39 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Mar 2005 11:27:36 -0500 16:27:36 -0000 j2BGRZx3025103 for ; Fri, 11 Mar 2005 11:27:35 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Mar 2005 11:27:35 -0500 (EST) Date: Fri, 11 Mar 2005 11:27:35 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANSLATION #1: MENUS AND DISPLAYS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4314 "All menu items are not always on display." (From an actual menu.) Correct translation: -Ax[Menu(x) > Display(x)] where: [[Display(x)]] = [[x]] is always on display. [[Menu(x)]] = [[x]] is a menu item. One student's mistranslation: | This is the only one I have questions about. My answer makes much more sense | to me than yours does. I have reasoned it out many times. The other ones I | understand why I was wrong, but this one you failed to convince me on. I | think this one may become your favorite, since a lot of people got it | "wrong" and will probably ask you about it: | | "All menu items are not always on display." | | I wrote: All(X) [Menu(X) => ~Display(X)], which to me means: "For | everything, if it is a menu item, then it is not the case that it is always | on display." | | You wrote: ~All(X) [Menu(X) => Display(X)], which is your answer, which to | me means: "It is not the case that for everything that is a menu item, it | follows that they are always on display." | | The negation of "always on display" led me to put the "not" in front of | "Display(X)". I believe that what I did is correct. ~Display(X) follows from | Menu(X). | | If you had put: "Not all menu items are always on display."... I would have | put your answer, or maybe: Exists(X) [Menu(X) ^ ~Display(X)], which I | believe is equivalent to the "Not all [...]" statement, if Aristotelian | logic is applicable and my mind is working well this early in the morning. | That sentence, however, is not equivalent to the one you gave us on the | question. | | Sorry if I seem to be doing this like a computer would... it makes the most | sense to me that way. Humans are not meant to do logic manually. I agree | with your statement about computers being much better at symbol | manipulation. | | If I did make a mistake, let me know where I messed up. It will probably | blow my mind. Get ready to have your mind blown :-) I agree that this one is very tricky, but, as Albert pointed out to me, I have already given you all an extra 2 points on this problem :-) Let's think about it slowly: 1. Consider a slightly simpler sentence: All menu items are not on display. Let [[D(x)]] = [[x]] is on display. I maintain that this sentence has the same syntactic structure as "All that glistens is not gold", namely: -Ax[Glistens(x) > Gold(x)] and NOT: Ax[Glistens(x) > -Gold(x)] So, for my simpler menu/display sentence above, we would have: -Ax[Menu(x) > D(x)] and NOT: Ax[Menu(x) > -D(x)] 2. Now, what about that pesky "always"? Let me propose the following representation of "x is always on display", where: [[D2(x,t)]] = [[x]] is on display at time [[t]] Using this, we would have: [[At.D2(x,t)]] = for all times [[t]], [[x]] is on display at time [[t]] = [[x]] is always on display. 3. Note that my predicate "D" above can be *reinterpreted* to mean "D2" as follows: Since "At.D2(x,t)" only has one free variable (the occurrences of "t" are all bound occurrences), it is equivalent to a predicate that has only one free variable; call this new predicate "D". (I.e., this new "D" encodes the "At.D2(_,t)" portion of "At.D2(x,t)".) In this case, "All menu items are not always on display" can be seen to have the same syntactic structure as "All menu items are not on display" and "All that glistens is not gold". If you prefer, you can think of "All menu items are not always on display" as "All menu items are not alwaysondisplay", where "alwaysondisplay" is represented as "At.D2(x,t)", or just "D(x)", using my new predicate introduced above. 4. So, "All menu items are not always on display" could be represented as: -Ax[Menu(x) > At.D2(x,t)] or as: -Ax[Menu(x) > D(x)] or, using the syntax required on the midterm, as: -Ax[Menu(x) > Display(x)]. 5. Having said all that, you might still not be convinced. So let me try another way (there's a 3rd way that I'll get to at the end if this second one doesn't work :-): You might want to paraphrase "All menu items are not always on display" as either: All menu items are such that it is not always the case that they are on display. or as: All menu items are such that it is not the case that they are always on display. The first of these would, indeed, be represented as: Ax[Menu(x) > -At.D2(x,t)] which is very close to the representation for the exam question that I marked as wrong. The second of the above paraphrases would *also* be represented that way! BUT: I maintain that these are NOT CORRECT paraphrases! To see why, here's my third explanation.. 6. What is the author of the menu trying to say? (And let's all admit that it's a really awkward, logically confusing way to say it!) I submit that they were trying to describe the following situation: Menu Display ---- ------- o o o o o o o Each circle represents an item on the menu (first column) or an item on display (2nd col.). Note that not all menu items are on display. That is, all menu items are not always on display. This situation is only correctly represented as: -Ax[Menu(x) > Display(x)] The wff: Ax[Menu(x) > -Display(x)] would represent this situation: Menu Display ---- ------- o o o o o And I don't think that that's what the menu writer intended. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Mar 11 11:38:24 2005 for ; Fri, 11 Mar 2005 11:38:24 -0500 (EST) for ; Fri, 11 Mar 2005 11:38:24 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Mar 2005 11:36:30 -0500 16:36:29 -0000 j2BGaTx3025172 for ; Fri, 11 Mar 2005 11:36:29 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Mar 2005 11:36:29 -0500 (EST) Date: Fri, 11 Mar 2005 11:36:29 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANSLATION #2: ROPERS AND ROPEES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4318 "When someone ropes someone, that someone can't move" Correct translation: AxAy[Ropes(x,y) > -Move(y)] where: [[Move(x)]] = [[x]] can move. [[Ropes(x,y)]] = [[x]] ropes [[y]] One student's mistranslation: | Ax Ey [Ropes(x,y) > -Move(y)] Call x the roper; call y the ropee (i.e., the one who is roped). What the correct wff says (which is a translation of the English sentence) is that for any roper and any ropee, if the roper ropes the ropee, then the ropee can't move. What the mistranslation says is: For any roper, there is a *particular* ropee who can't move if he is roped. But the English sentence doesn't really mean (despite the correct use of "someone"!) that only a particular ropee can't move; rather, it means that *anyone* who the roper ropes can't move. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Mar 11 11:50:20 2005 for ; Fri, 11 Mar 2005 11:50:20 -0500 (EST) for ; Fri, 11 Mar 2005 11:50:20 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Mar 2005 11:49:52 -0500 16:49:52 -0000 j2BGnpx3025284 for ; Fri, 11 Mar 2005 11:49:51 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Mar 2005 11:49:51 -0500 (EST) Date: Fri, 11 Mar 2005 11:49:51 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE/4/563: MISTRANSLATION #3: MAPS AND COUNTRIES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4320 ------------------------------------------------------------------------ Subject: CSE/4/563: MISTRANSLATION #3: MAPS AND COUNTRIES ------------------------------------------------------------------------ "The United States is a country. So, a map of the United States is a map of a country!" Correct translation: Country(us) |- Ax[Map(x,us) > Ey[Country(y) ^ Map(x,y)]] or: Country(us) |- Ax[x=map(us) > Ey[Country(y) ^ x=map(y)]] where: [[us]] = the United States (a constant function-symbol or term) [[map(x)]] = a map of [[x]] (a 1-place function symbol) [[Country(x)]] = [[x]] is a country. (a 1-place predicate) [[Map(x,y)]] = [[x]] is a map of [[y]]. (a 2-place predicate) Note that the turnstile could also have been a ">" (depending on whether you interpret "So" as "therefore", which would be the turnstile, or as "then", which would be a ">". One student's mistranslation: Country(us) ^ AxEy[(y = us) ^ Map(y) > Country(us) ^ AxEy[(Country(y) ^ Map(x,Map(us))) > Map(x,y)]] First mistake: The "So" is represented as a conjunction. This is not too bad, since the English does assert that the US is a country. Perhaps a better way (for both the student and I) to have represented this would have been: Country(us) ^ (Country(us) > ...) Now, consider the student's 2nd conjunct: x ranges over maps, and y ranges over countries. So the 2nd conjunct says: For each map x, there is a country y such that: y is the United States and ... and what? Map(y) is syntactically incorrect! Capital-M "Map" is a 2-place predicate, but here it's used as a 1-place predicate. Lower-case-m "map" is a 1-place function symbol, but if that's what the student intended to use here, it's still incorrect, since "map(y)" is not a wff, hence can't be the antecedent of a conditional. I won't try to go beyond this error. If you would like to fix this syntactic error and submit another translation for my consideration, please go ahead! | | 7.b) "No Two Things are the same" | | Ax.x ^ Ay.y ^ -(x=y) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Mar 11 15:43:59 2005 for ; Fri, 11 Mar 2005 15:43:58 -0500 (EST) for ; Fri, 11 Mar 2005 15:43:58 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Mar 2005 15:43:06 -0500 20:43:00 -0000 j2BKh0x3028855 for ; Fri, 11 Mar 2005 15:43:00 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Mar 2005 15:43:00 -0500 (EST) Date: Fri, 11 Mar 2005 15:43:00 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANSATIONS #4: TWO THINGS THE SAME? To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4350 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANSATIONS #4: TWO THINGS THE SAME? ------------------------------------------------------------------------ "No two things are the same" Correct translation (given our limited semantics): AxAy.-(x=y) or: AxAy[-(x=y) > -(x=y)] (Please see the midterm answers for more discussion.) One student's mistranslation: Ax.x ^ Ay.y ^ -(x=y) This is syntactically incorrect, since the first conjunct (and the second) are syntactically incorrect: What follows a quantifier+variable phrase must be a wff. But "x" is a variable, not a wff (ditto for "y" in the second conjunct). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Mar 11 15:52:57 2005 for ; Fri, 11 Mar 2005 15:52:56 -0500 (EST) for ; Fri, 11 Mar 2005 15:52:56 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 11 Mar 2005 15:52:13 -0500 20:52:13 -0000 j2BKqDx3028972 for ; Fri, 11 Mar 2005 15:52:13 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Fri, 11 Mar 2005 15:52:13 -0500 (EST) Date: Fri, 11 Mar 2005 15:52:13 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANSLATION #5: GREAT LAKES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4353 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANSLATION #5: GREAT LAKES ------------------------------------------------------------------------ "All the Great Lakes won't freeze, but all is not well." (Heard on NPR.) Correct translation: -Ax[GreatLake(x) > Freeze(x)] ^ -Ax.Well(x) where: [[Freeze(x)]] = [[x]] will freeze. [[GreatLake(x)]] = [[x]] is a Great Lake. [[Well(x)]] = [[x]] is well. One student's mistranslation: Ax[GreatLake(x) > -Freeze(x)] ^ Ay[-Well(y)] Let's look at each conjunct separately. Left conjunct: This says: "for any x, if x is a Great Lake, then x won't freeze" But, as with "All that glitters is not gold", the original English sentence just says that not all of the Great Lakes will freeze. If you're still not convinced, consider this: "All dogs aren't pets". Surely this doesn't mean that no dogs are pets. It can only mean that some dogs aren't pets. Right conjunct: This says: "for any y, it is not well" Again, the original English is an idiomatic way of saying that something is not well, not that everything isn't well. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Mar 12 13:32:41 2005 for ; Sat, 12 Mar 2005 13:32:40 -0500 (EST) for ; Sat, 12 Mar 2005 13:32:40 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 12 Mar 2005 13:31:52 -0500 18:31:52 -0000 for ; Sat, 12 Mar 2005 13:31:52 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sat, 12 Mar 2005 13:31:52 -0500 (EST) Date: Sat, 12 Mar 2005 13:31:52 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 7 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4449 ------------------------------------------------------------------------ Subject: CSE 4/563: HW 7 QUESTION ------------------------------------------------------------------------ A student writes: | I have two relatively straightforward hw questions | concerning question 29:3: | | 1. B & L introduce e s.t. [[Elt(e, x)]] = e is an element | of x. Now e is not one of the variables in our language, but | this is seems to be unequivocably a variable, am I mistaken? | The reason I ask is that I wish to quantify over it. Good point! It is not an officially legal variable of B&L's language, but it *is* of ours, since I allow computer-programming-language-style identifiers among the variables. So, yes, treat "e" as a variable. Note to readers: In the next question, below, I've omitted the student's answer to one of the HW problems, which was part of the student's question, and have put a substitute that makes the same point in brackets: | 2. More interestingly, B & L give u(x, y) to be the union of two | sets, which of course will mathematically itself be a set. For the | last translation in 3, the only way I see to represent this (granted, I | have only spent 10-20 mintues thinking about it) is to [use "Elt(e, u(x,y)"] | [O]f course embedded predicates are not allowed in FOPC, | but isn't this a special instance given u(x,y) reduces to a single | dimension of the same domain as x and y. Therefore, this is not the | same issue as the exam question where essentially | Says(x, ugly(w)) | right? Right, but not quite for the reason you seem to suggest. "u" is not a predicate; it is a 2-place function symbol. Thus, it is perfectly legal for it to appear in argument position within a predicate formula. I.e., you *can* embed function symbols (but you cannot embed predicates). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 14:49:05 2005 for ; Sun, 13 Mar 2005 14:49:04 -0500 (EST) for ; Sun, 13 Mar 2005 14:49:04 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 14:48:19 -0500 19:48:19 -0000 for ; Sun, 13 Mar 2005 14:48:18 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 14:48:18 -0500 (EST) Date: Sun, 13 Mar 2005 14:48:18 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PRAGMATICS: the missing lecture To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4492 ------------------------------------------------------------------------ Subject: CSE 4/563: PRAGMATICS: the missing lecture ------------------------------------------------------------------------ I mentioned in lecture on Friday that I would not cover the B&L section on "pragmatics", in order to be able to get to automated theorem proving after the break. So, I've put a version of the lecture I would have given on the website: http://www.cse.buffalo.edu/~rapaport/563S05/pragmatics.pdf From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 17:01:18 2005 for ; Sun, 13 Mar 2005 17:01:18 -0500 (EST) for ; Sun, 13 Mar 2005 17:01:17 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 17:00:02 -0500 22:00:02 -0000 for ; Sun, 13 Mar 2005 17:00:00 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 17:00:00 -0500 (EST) Date: Sun, 13 Mar 2005 17:00:00 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANSLATION #6: ROPER & ROPEE, PART 2 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: Junk X-UID: 4496 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANSLATION #6: ROPER & ROPEE, PART 2 ------------------------------------------------------------------------ "When someone ropes someone, that someone can't move" Correct translation: AxAy[Ropes(x,y) > -Move(y)] where: [[Move(x)]] = [[x]] can move. [[Ropes(x,y)]] = [[x]] ropes [[y]] Another student's mistranslation: ExEy[Ropes(x,y) ^ -Move(y) ^ -(x=y)] (the -(x=y) was added assuming that someone could not rope themselves) The main problem with this is that it asserts that there is someone who ropes and there is someone who gets roped (i.e., that there is a roper and that there is a ropee), but the original English doesn't imply that (two) such individuals exist, merely that, if they *did* exist, then the second wouldn't be able to move. The 3rd conjunct goes beyond the original English, and makes an assumption about the distinctness of the roper and the ropee. It's a normal assumption, but imagine a clumsy roper, or a clown at a rodeo, who ropes himself. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 17:09:01 2005 for ; Sun, 13 Mar 2005 17:09:00 -0500 (EST) for ; Sun, 13 Mar 2005 17:09:00 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 17:08:22 -0500 22:08:22 -0000 for ; Sun, 13 Mar 2005 17:08:21 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 17:08:21 -0500 (EST) Date: Sun, 13 Mar 2005 17:08:21 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 7: MAPS AND COUNTRIES, part 2 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4497 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 7: MAPS AND COUNTRIES, part 2 ------------------------------------------------------------------------ "The United States is a country. So, a map of the United States is a map of a country!" Correct translation: Country(us) |- Ax[Map(x,us) > Ey[Country(y) ^ Map(x,y)]] or: Country(us) |- Ax[x=map(us) > Ey[Country(y) ^ x=map(y)]] where: [[us]] = the United States (a constant function-symbol or term) [[map(x)]] = a map of [[x]] (a 1-place function symbol) [[Country(x)]] = [[x]] is a country. (a 1-place predicate) [[Map(x,y)]] = [[x]] is a map of [[y]]. (a 2-place predicate) Note that the turnstile could also have been a ">" (depending on whether you interpret "So" as "therefore", which would be the turnstile, or as "then", which would be a ">". Another student's mistranslation: Country(us) > (Ex[(Country(x) ^ map(us)) > map(x)) One problem with this is that the parentheses at the end are mismatched; that might just be an email-typo. I'll assume that what was intended was: Country(us) > Ex[(Country(x) ^ map(us)) > map(x)] Even so, this makes an error similar to the one discussed earlier: "map" is a 1-place function symbol; hence, it cannot appear as a conjunct (it doesn't express a sentence; it's neither true nor false). Translated back into English, the above wff literally says: "If the US is a country, then there is an x such that, if x is a country and a map of the US, then a map of x" which is just ungrammatical. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 19:45:06 2005 for ; Sun, 13 Mar 2005 19:45:06 -0500 (EST) for ; Sun, 13 Mar 2005 19:45:05 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 19:44:12 -0500 00:44:12 -0000 for ; Sun, 13 Mar 2005 19:44:11 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 19:44:11 -0500 (EST) Date: Sun, 13 Mar 2005 19:44:11 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 8: BUGS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4510 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 8: BUGS ------------------------------------------------------------------------ "Some bugs are dead. And some are alive" Correct translation: Ex[Bug(x) ^ Dead(x)] ^ Ey[Bug(y) ^ Alive(y)] where: [[Alive(x)]] = [[x]] is alive. [[Bug(x)]] = [[x] is a bug. [[Dead(x)]] = [[x]] is dead. One student's mistranslation: | Ex[bug(x) ^ dead(x) ) ^ live(x)] | mistake: assumed since talking about bugs and x is a bug that Ex will | translate to some bugs are dead and alive, but I think it could of been | right saying Ex[bug(x) ^ dead(x) v live(x)] as an alternative to the | solution. (I assume that the student meant to say "it could *have* been right...".) There's a stray ")" just before the 2nd "^"; I'll assume that's a typo. (If not, then there needs to be a matching "(" before "bug", which, by the way, should have been "Bug" with a capital "B", and the other predicates should have been caps, too. This may seem picky (and it would have been if I had taken points off for it), but note that in the country/map sentence, I carefully distinguished between "Map" and "map"!) And "live" should have been "Alive" :-) The real problem, however, is that this says that some bugs are dead and alive. But that would be true if there were a single bug that was (somehow) both. The English sentence, however, is true if there are at least 2 bugs, one dead, and one alive. More to the point, *both* would be true were there a single dead-and-alive bug, but the mistranslation would be *false* if there were only 2 bugs, one dead, one alive, whereas the English would be true in that circumstance.. What about: Ex[bug(x) ^ dead(x) v live(x)] This is *syntactically* incorrect, since it's ambiguous between these two logically equivalent wffs: Ex[(Bug(x) ^ Dead(x)) v Alive(x)] (1) and Ex[Bug(x) ^ (Dead(x) v Alive(x))] (2) (1) says that there is something that is either a dead bug or alive (but, if it's alive, it doesn't say whether it's a bug or not). The English, however, requires that some *bugs* be alive. Side issue: Does it really? Well, I agree that it doesn't literally say that some *bugs* are alive. But, in ordinary, everyday discourse, that surely is what it is intended to mean. (2) says that there's a bug, and either it's dead or it's alive, which, given some reasonable assumptions about the ordinary English meanings of "Dead" and "Alive", is a tautology, which the English sentence isn't. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 20:00:46 2005 for ; Sun, 13 Mar 2005 20:00:45 -0500 (EST) for ; Sun, 13 Mar 2005 20:00:45 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 19:57:03 -0500 00:55:44 -0000 for ; Sun, 13 Mar 2005 19:55:44 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 19:55:44 -0500 (EST) Date: Sun, 13 Mar 2005 19:55:44 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 9: countries & maps again To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 9: countries & maps again ------------------------------------------------------------------------ ""The United States is a country. So, a map of the United States is a map of a country!" Correct translation: Country(us) |- Ax[Map(x,us) > Ey[Country(y) ^ Map(x,y)]] or: Country(us) |- Ax[x=map(us) > Ey[Country(y) ^ x=map(y)]] Here are a couple more mistranslations: (1) Country(us) ^ Ax[(Map(x) ^ x=us) > Country(x)] (2) Country(us) > Ax[Map(x,us) > Ay[Map(x,y) > Country(y)]] (1) makes a syntactic error: "Map", in my semantics, is a 2-place predicate; here, it is used as a 1-place predicate. Or perhaps the student meant it as my 1-place function symbol; but then it can't appear as a conjunct. (2) says that, if the US is a country, then all maps of the US have the following property: for all y, if x (which, remember, is any map of the US) is a map of y (i.e., of *any* y), then y is a country; i.e., all things that x is a map of are countries. Now, this is true. But it's not what the English sentence says. The "So" part of the English tells us something about maps of the US, namely, that they are maps of countries. (2) doesn't tell us anything about maps per se; rather, it gives us a sufficient condition for something to be a country! From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 20:05:51 2005 for ; Sun, 13 Mar 2005 20:05:51 -0500 (EST) for ; Sun, 13 Mar 2005 20:05:50 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 20:02:50 -0500 01:02:50 -0000 for ; Sun, 13 Mar 2005 20:02:50 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 20:02:49 -0500 (EST) Date: Sun, 13 Mar 2005 20:02:49 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 10 (I think): Great Lakes again To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 10 (I think): Great Lakes again ------------------------------------------------------------------------ "All the Great Lakes won't freeze, but all is not well." (Heard on NPR.) Correct translation: -Ax[GreatLake(x) > Freeze(x)] ^ -Ay.Well(y) One more mistranslation: (1) | ~Ax[greatlakes(x) > Ey[freeze(y) > Az[well(z)]]] | problem translating "but" and also knowing when to use 1 variable verses | 2 variables, especially in problems that require 2 or more quantifiers Ignoring what I'll take to be typos (lowercase predicates), (1) says that it's not the case that, for all x, if x is a Great Lake, then something (not necessarily a Great Lake) is such that if it freezes, then all *is* well. I hope you see that that's not what the English says. "But" is always translated in FOL as "^". To decide how many variables to use, figure out how many things you're talking about. In the first conjunct of the English sentence, there's only one thing: a Great Lake. In the second conjunct, there's another thing: things that are (not) well (which need not be Great Lakes). From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 20:30:22 2005 for ; Sun, 13 Mar 2005 20:30:22 -0500 (EST) for ; Sun, 13 Mar 2005 20:30:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 20:29:56 -0500 01:29:56 -0000 for ; Sun, 13 Mar 2005 20:29:55 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 20:29:55 -0500 (EST) Date: Sun, 13 Mar 2005 20:29:55 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 11: Menus again To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4512 ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 11: Menus again ------------------------------------------------------------------------ "All menu items are not always on display." (From an actual menu.) Correct translation: -Ax[Menu(x) > Display(x)] Here are three other mistranslations: | ~Ax[menu(x) > Ey[Display(y)]] (1) | again same mistake should of either replaced y with x or did | ~Ax[menu(x) > Ey[menu(y) ^ y=x ^ display(y)]] (2) | Also problem translating always | However, I still think that my solution for Menu & Display question is | also correct: | | Ex[Menu(x) ^ Display(x)] (3) | | Why isn't this one correct? Because what I am trying to say is that | there are some menu items that are always on display which is what the | sentence implies. (1) says that it's not the case that, for all x, if x is a menu item, then there is something (not necessarily a menu item) always on display. But even if we all disagree about exactly what the English means, surely it means that the things that are (not) on display are menu items, not anything whatsoever. (And "of" should be "have" in the student's comment!!) Note that replacing "y" with "x" won't fix it; that would just make it: -Ax[Menu(x) > Ex.Display(x)] but both "x" and "y" are bound variables, so they are interchangeable without changing the meaning. What about (2)? As with (1), it seems to introduce a new menu item (since it quantifies over "y", not "x"). But it then equates that "y" with "x". So those two conjuncts are eliminable. So, (2) is logically equivalent to: -Ax[Menu(x) > Display(x)] and this, I have already argued, *is* correct! (3), however, isn't. It indeed says that some menu items are always on display. But that's *not* what the English says. It doesn't even imply it. (But I'll leave the proof of that as an exercise :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 13 21:52:15 2005 for ; Sun, 13 Mar 2005 21:52:15 -0500 (EST) for ; Sun, 13 Mar 2005 21:52:15 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 13 Mar 2005 21:51:50 -0500 02:51:50 -0000 for ; Sun, 13 Mar 2005 21:51:49 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 13 Mar 2005 21:51:49 -0500 (EST) Date: Sun, 13 Mar 2005 21:51:49 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MISTRANS 12: Roping again To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MISTRANS 12: Roping again ------------------------------------------------------------------------ "When someone ropes someone, that someone can't move" Correct translation: AxAy[Ropes(x,y) > -Move(y)] Here is another incorrect translation: ExEy[someone(x) ^ someone(y) ^ rope(x,y) > ~move(y)] There are a number of problems with this. First, technically it's syntactically incorrect, since it's not clear what the antecedent of "-move(y)" is: Is it just "rope(x,y)"? Or is it all 3 of the conjuncts? Or just the last 2 of them? I'll assume that it's all 3. Second, "someone" is not one of the legal predicates that I allowed you to use. Using it is a bit like speaking a language and occasionally throwing in a new word that no one knows or that comes from another language. In real life, you can often figure out what such a word means from context, but in KRR it's not allowed. If we fix these problems (and capitalize the other predicates :-), we get a wff that looks like mine except for using existential quantifiers where I used universals: ExEy[Ropes(x,y) > -Move(y)] This says that there are two individuals that are such that, *if* the first ropes the second, then the second can't move. Suppose that there are two individuals such that neither ropes the other. Then this formula is true, but only for those two guys. The original English sentence holds true for anyone. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 14 14:43:31 2005 for ; Mon, 14 Mar 2005 14:43:30 -0500 (EST) for ; Mon, 14 Mar 2005 14:43:30 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 14 Mar 2005 14:43:24 -0500 19:43:24 -0000 for ; Mon, 14 Mar 2005 14:43:23 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 14 Mar 2005 14:43:23 -0500 (EST) Date: Mon, 14 Mar 2005 14:43:23 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PROJECT 1 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROJECT 1 QUESTION ------------------------------------------------------------------------ A student writes: | The clause form algorithm you prepared for us does not tell us what to | do with unbound variables (for example "Ex( (f(x,a)=>y) or (y) )" ). | In such cases, I am not sure if to (may be) add a "for all y" before the | "there exists". This is left unspecified, since it will never occur in practice (a KRR system only works with sentences, i.e., wffs with no free variables). Thus, you may either ignore this case, or else deal with it any way you want. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 16 11:56:22 2005 for ; Wed, 16 Mar 2005 11:56:22 -0500 (EST) for ; Wed, 16 Mar 2005 11:56:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 16 Mar 2005 11:49:54 -0500 16:49:54 -0000 for ; Wed, 16 Mar 2005 11:49:54 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 16 Mar 2005 11:49:54 -0500 (EST) Date: Wed, 16 Mar 2005 11:49:54 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 7 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4687 ------------------------------------------------------------------------ Subject: CSE 4/563: HW 7 QUESTION ------------------------------------------------------------------------ Question from ctclader@cse.Buffalo.EDU Wed Mar 16 10:51:34 2005 | I have a quick question concerning problem 1 on hw 7. | | Say I have the wff VxVy[(P(x,y) ^ P(y,x)) > P(a,x)] It's not a good idea to use "V" as the universal quantifier, since another FOL notation that's fairly common uses it for the *existential* quantifier! (I mentioned this in class on Friday; recall that this alternative notation scheme uses a "capital ^" (i.e., an inverted capital "V") for the universal quantifier.) | If I make it so that it is impossible, for all y and x, for both | P(x,y) and P(y,x) to be true, that I can then assume that this entire | sentence is true, since F > T is always T? | | For example let D = All the people in the world. | Let [[P(x,y)]] be "x is the father of y" | Let [[a]] be "albert" | | Thus it is impossible for both P(x,y) and P(y,x) to be true, since | no one is the father of themselves. | | So the sentence more or less translates to: For all things such that | x is a person and y is a person, if x is the father of y, and y is | the father of x, then albert is the father of x. | | Can I take this sentence to be true? In my opinion it is, since | you can't possibly come up with a counter-example. | | Thanks for your help, | Chris Answer from ag33@cse.Buffalo.EDU Wed Mar 16 11:07:39 2005 | Your analysis is correct...the statement will be trivially true under | this interpretation since the antecedent is always false. Just be | sure that the semantics of the interpretation are the same for the | other two wffs including the one you are trying to make false | (i.e., remember that P(x,y) still means "x is the father of y" and a | still means "albert" in the other two wffs). | | Albert | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 16 14:58:22 2005 for ; Wed, 16 Mar 2005 14:58:22 -0500 (EST) for ; Wed, 16 Mar 2005 14:58:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 16 Mar 2005 14:57:19 -0500 19:57:16 -0000 for ; Wed, 16 Mar 2005 14:57:15 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 16 Mar 2005 14:57:15 -0500 (EST) Date: Wed, 16 Mar 2005 14:57:15 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: OPEN MIND KRR To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: OPEN MIND KRR ------------------------------------------------------------------------ You might have some fun with this. It's a serious research project in KRR/ontology: | The webaddress for Open mind common sense project at the Honda Research | Institute is: | | http://openmind.hri-us.com/ | | You will have to register (provide user information) if you want to see | the kind of data they collect and the sentence templates they use for | collecting this common sense data. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 17 00:28:23 2005 for ; Thu, 17 Mar 2005 00:28:23 -0500 (EST) for ; Thu, 17 Mar 2005 00:28:22 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 17 Mar 2005 00:27:37 -0500 05:27:37 -0000 for ; Thu, 17 Mar 2005 00:27:37 -0500 (EST) -0500 (EST) Date: Thu, 17 Mar 2005 00:27:36 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: What is the "one" in everyone? To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: NotJunk X-UID: 4729 Hi all, I finished reading Gottlob Frege's "The Foundations of Arithmetic" this break (for my own research), and he has a few words to say about why a sentence like "Everyone is happy" is translated as Ax(Happy(x)) in FOL...and NOT translated as Ax(One(x) > Happy(x)) "In isolation, however, it seems that 'one' cannot be a predicate. This is even clearer if we take the plural. Whereas we can combine 'Solon was wise' and 'Thales was wise' into 'Solon and Thales were wise', we cannot say 'Solon and Thales were one'. But it is hard to see why this should be impossible if 'one' were a property of both Solon and of Thales in the same way that 'wise' is" (40-41) he then says, with some frustration: "How can it makes sense to ascribe the property 'one' to any object whatever, when every object, according as to how we look at it, can be either one or not one? How can a science which bases its claim to fame precisely on being as definite and accurate as possible repose on a concept as hazy as this?" (41) And this is all before Godel showed up and demonstrated that his "science" (i.e., arithmetic) was on far shakier ground... ouch! Frege also questions whether or not 'one' changes the semantics of a sentence that would otherwise use the indefinite article, 'a'. Consider how you would represent the following: (a) One bird flew through the window (b) A bird flew throught the window Do (a) and (b) mean the same thing? Hope you are having a good break, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Mar 19 06:46:38 2005 for ; Sat, 19 Mar 2005 06:46:38 -0500 (EST) for ; Sat, 19 Mar 2005 06:46:38 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 19 Mar 2005 06:45:48 -0500 11:45:47 -0000 for ; Sat, 19 Mar 2005 06:45:47 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sat, 19 Mar 2005 06:45:47 -0500 (EST) Date: Sat, 19 Mar 2005 06:45:47 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUESTION RE: LISP VS. JAVA To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: QUESTION RE: LISP VS. JAVA ------------------------------------------------------------------------ A student writes: | I was just wondering... I know what LISP is, and I'm familiar with the | basic syntax, but I haven't touched it in a few semesters and I don't | remember ever being very fluent with it. I would definately need to | re-learn a few things about it, and get comfortable with it again. | | Java, on the other hand, is something I'm very good with. | | Given the lack of difficulty of part one, I don't see a problem using | either language, however, the later parts may make me wish that I | chose one way over the other (LISP for its functionality or Java for | my experience with it). | | The question is, which do I chance it on? Could you give me a | recommendation from this perspective? Or perhaps tel me what features | of LISP make it more efficient/useful for the project that Java lacks. Not an easy question to answer, since there are practical considerations as well as theoretical ones. Given the very short, finite amount of time you have for completing the project, my recommendation would be to go with the language you are more familiar with, and try to solve the problems in that more-or-less familiar domain. It *can* be done (since all programming languages are Turing-equivalent). Since I'm not familiar with Java, I can't really tell you what (dis)advantages it might have. For more general advice on the *advantages* of Lisp, take a look at some of the links on "Good Reasons to Learn Lisp", at: http://www.cse.buffalo.edu/~rapaport/lisp.html (or see the link to that page from the syllabus), especially Paul Graham's "Beating the Averages". From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 20 11:34:00 2005 for ; Sun, 20 Mar 2005 11:33:59 -0500 (EST) for ; Sun, 20 Mar 2005 11:33:59 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Mar 2005 11:33:35 -0500 16:23:33 -0000 Date: Sun, 20 Mar 2005 11:23:33 -0500 Reply-To: Joaquin Carbonara From: Joaquin Carbonara Subject: Project 3(b) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU I quote the assignment "explain the relationship between Moore's example and this problem". The example in Moore is a nice problem by itself, but also part of a larger discussion. Do we need to address both, the problem in itself and the larger issue or just the example (of course, as they relate to project part 3(b))? Joaquin From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 20 11:44:55 2005 for ; Sun, 20 Mar 2005 11:44:55 -0500 (EST) for ; Sun, 20 Mar 2005 11:44:55 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Mar 2005 11:41:46 -0500 16:41:46 -0000 j2KGfjBt003801; Sun, 20 Mar 2005 11:41:45 -0500 (EST) -0500 (EST) Date: Sun, 20 Mar 2005 11:41:45 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: Project 3(b) Comments: To: jc39@BUFFALO.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 4956 Joaquin writes: | I quote the assignment "explain the relationship between Moore's example | and this problem". | | The example in Moore is a nice problem by itself, but also part of a | larger discussion. Do we need to address both, the problem in itself and | the larger issue or just the example (of course, as they relate to project | part 3(b))? As you wish :-) At a minimum, just the example. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 20 15:36:44 2005 for ; Sun, 20 Mar 2005 15:36:44 -0500 (EST) for ; Sun, 20 Mar 2005 15:36:44 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 20 Mar 2005 15:36:23 -0500 20:36:23 -0000 for ; Sun, 20 Mar 2005 15:36:22 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 20 Mar 2005 15:36:22 -0500 (EST) Date: Sun, 20 Mar 2005 15:36:22 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUESTION ABOUT CONNECTIVE PRECEDENCE ORDERING To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: QUESTION ABOUT CONNECTIVE PRECEDENCE ORDERING ------------------------------------------------------------------------ A student writes: | > For the fol connectives, its this the correct order (and only symbols)? | > | > highest not | > conjunction | > disjunction | > implication | > lowest if and only if and Albert responded: | You may want to ask Dr. Rapaport about the official order of precedence. | I believe that, for our version of FOL, since you technically *have to* | introduce parentheses whenever you introduce a binary connective, there | should be no ambiguity of operator precedence. Albert is correct. The general answer is this: There is no standard ordering. Every author is allowed to introduce their own ordering as long as it is unambiguous. Since this is the case, and since I, personally, have never been able to remember the ordering, I never use one, but always insist on full parenthesization except where there is no possibility of confusion (e.g., outermost parentheses can be omitted). Our text's official ordering is listed on pp. 17-18. You may use any ordering you wish, as long as you make clear what it is. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 21 10:02:02 2005 for ; Mon, 21 Mar 2005 10:02:01 -0500 (EST) for ; Mon, 21 Mar 2005 10:02:01 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 21 Mar 2005 09:59:32 -0500 14:59:32 -0000 j2LExVso010389 for ; Mon, 21 Mar 2005 09:59:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 21 Mar 2005 09:59:31 -0500 (EST) Date: Mon, 21 Mar 2005 09:59:31 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: FOL INFERENCE RULES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 5019 ------------------------------------------------------------------------ Subject: CSE 4/563: FOL INFERENCE RULES ------------------------------------------------------------------------ Instead of spending valuable lecture time on this topic, some of which has been covered in recitation anyway, here, for future reference, are the... Rules of Inference for Our FOL Natural-Deduction System ======================================================================== All rules of inference for our propositional-logic natural deduction system, plus the following: Notation: Let P be a wff (in Latex, I'd use \alpha :-) Let t be a constant or variable term Then write "P(t)" for: any wff (whose predicate is "P") containing 0 or more occurrences of t Note: "P(t)" could be something like: AxEy.P(x,t,y) So, "At.P(t)" could be: AtAxEy.P(x,t,y) Now, here are the new rules of inference: ------------------------------------------------------------------------ A Intro: From P(s), where s is an "arbitrary" individual about which no special assumptions are made (e.g., a triangle, but neither equilateral, isosceles, or scalene, even though any triangle must be one of these!) ------------------------------------------------------ Infer Ax.P(x) As a strategy: To prove Ax.P(x) show P(s) for arbitrary s ------------------------------------------------------------------------ ------------------------------------------------------------------------ A Elim: From Ax.P(x) ----------------------------------- Infer P(t), where t can be any term As a strategy: To prove P(t), show Ax.P(x) ------------------------------------------------------------------------ ------------------------------------------------------------------------ E Intro: From P(t), where t is a specific individual ------------------------------------------- Infer Ex.P(x) As a strategy: To prove Ex.P(x) show P(t) for some t (i.e., find some t such that P(t) ) ------------------------------------------------------------------------ ------------------------------------------------------------------------ E Elim: From Ex.P(x) ----------------------------------------------------- Infer P(s), where s is a "new" individual about which no assumptions are made (e.g., suppose thatsomeone in this class failed the exam. Call that student Sandy, (assuming that there is no actual student in this class named "Sandy"; if there were, I'd have to use a different made-up name!) As a strategy: To prove P(s), for some s show Ex.P(x) Actually: This rule is usually used to get a P (i.e., to get an instance of something that has the property [[P]]) that can then be universally instantiated: From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 22 09:01:29 2005 for ; Tue, 22 Mar 2005 09:01:29 -0500 (EST) for ; Tue, 22 Mar 2005 09:01:28 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 22 Mar 2005 09:00:21 -0500 14:00:21 -0000 j2ME0Kso024409 for ; Tue, 22 Mar 2005 09:00:20 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 22 Mar 2005 09:00:20 -0500 (EST) Date: Tue, 22 Mar 2005 09:00:20 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PROJECT QUESTION ABOUT UNIFICATION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROJECT QUESTION ABOUT UNIFICATION ------------------------------------------------------------------------ A student writes: | I have a question on the unification part of the project. I think I | understand the idea of unification. Given a set of atoms, we determine | a variable substitution that will allow those atoms to "look the same" | when the substitutions are applied. Correct. | It seems clear that if the | predicates don't match, unification automatically fails. Since I | haven't looked at the resolution part of the project, I'm not certain | where this will plug in, and so I might be missing exactly what we'll | be operating on - for example, will the unifier ever be given a | negation of an atom in the set? I'm going to look at the book for more | examples, but all of the ones I've seen so far are just the two | predicate case. For purposes of automated theorem proving, you'll only ever be trying to unify two wffs, one of which will be the negation of the other (so, actually, you'll be trying to unify 2 wffs of the forms P(t1,...) and -P(t2,...) (where t1 and t2 are terms) | Now for the real question (since I've made assumptions in my code about | the above!). When you are doing unification, are you substituting only | using the term values (variables, functions, constants) in the supplied | set, or is it legal (and useful) to do substitutions using terms from | other clauses? For example, let's say I have this as my set of | clauses, after doing all of the sentence translations: | | { [ P(a,b,x) Q(x,y) ] [Q(a,c)] [P(z,w,f(y))] } | | So, the set of variables in the "world" is w,x,y,z, the constants a,b,c | and one function f(y). | | If I'm trying (for whatever reason) to unify { P(a,b,x), P(z,w,f(y) }, | can I introduce c? Is the answer "yes, but it won't help you get the | MGU?" Think of unification recursively. You can't unify 2 distinct constants. You can unify a constant and a variable by substituting the constant for the variable. You can unify a functional term--i.e., f(t1,...)--with a variable by substituting the functional term for the variable, but note that that functional term might contain occurrences of that same variable; that's a special case that has to be handled separately. You can unify two wffs by unifying each of their subwffs, subject to the above special problem. But, to make matters more complex, the notion of "substitution" has to be carefully defined. The book has all the details, and I will go over them in lecture. For a preview, take a look at "Notes on the Unification Algorithm" at: http://www.cse.buffalo.edu/~rapaport/663/F03/unification.pdf (also accessible from the new "FOL: Automated Theorem Proving" page, which is at: http://www.cse.buffalo.edu/~rapaport/563S05/fol-atp.html | If we're going to be getting to all of this soon, don't bother with | details, but please give me enough guidance to continue coding. :-) Hope this helps. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 23 12:22:56 2005 for ; Wed, 23 Mar 2005 12:22:56 -0500 (EST) for ; Wed, 23 Mar 2005 12:22:55 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 23 Mar 2005 12:21:45 -0500 17:21:45 -0000 j2NHLin9006388 for ; Wed, 23 Mar 2005 12:21:44 -0500 (EST) 12:21:44 -0500 (EST) Date: Wed, 23 Mar 2005 12:21:44 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: FOL syntactic proof To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 5236 Hello All, I wanted to cover clause form and project related issues in recitation this week...so I didn't get time to go over this proof below (which is used in clause form during the "push negation inward" phase). -ExF(x) |- Ax(-F(x)) 1. -ExF(x) :premise *2. F(a) :temp. assumption *3. ExF(x) :EIntro 2 *4. -ExF(x) :send 1 *5. -F(a) :-Intro 2,3,4 6. -F(a) :return 5 7. Ax(-F(a)) :AIntro 6 {verify for yourselves that this satisfies the 3 constraints on AIntro} Try to do -AxF(x) |- Ex(-F(x)) on your own...I believe it is a bit trickier :-) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 23 15:59:09 2005 for ; Wed, 23 Mar 2005 15:59:09 -0500 (EST) for ; Wed, 23 Mar 2005 15:59:09 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 23 Mar 2005 15:57:53 -0500 20:57:53 -0000 j2NKvrso007119 for ; Wed, 23 Mar 2005 15:57:53 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Wed, 23 Mar 2005 15:57:53 -0500 (EST) Date: Wed, 23 Mar 2005 15:57:53 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 7 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 7 QUESTION ------------------------------------------------------------------------ A student writes: | I have a question regarding questions 3(b) and 4 from homework 7: | | Last week in lecture, you gave an example of a proof in the form: | | I,u |= - Ax.alpha iff I,u |= Ex[-alpha] | | etc. | | Should the proofs for questions 3 and 4 be in this form, or should they be | syntactic proofs in FOL similar to the syntactic proofs in propositional | logic we were doing before? As the text says, "show using ... interpretations". Interpretations are semantic beasts, not syntactic, so you need to use |=, not |-. I.e., the proofs should be in the form above, not syntactic proofs in FOL. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 23 18:32:29 2005 for ; Wed, 23 Mar 2005 18:32:29 -0500 (EST) for ; Wed, 23 Mar 2005 18:32:29 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 23 Mar 2005 18:31:43 -0500 23:31:43 -0000 X-Originating-IP: 128.205.245.49 Date: Wed, 23 Mar 2005 18:31:43 -0500 Reply-To: Rahul Krishna From: Rahul Krishna Subject: Skolemization To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: Junk X-UID: 5270 I have a question in the Skolemization step of the conversion to Clause Form algorithm. Let us consider the FOL wff: Ax Ay [P(x) ^ Q(y)] v Ez.R(z) .....(1) Moving the quantifiers to the left, we have Ax Ay Ez [P(x) ^ Q(y)] v R(z) >From step 2(a), this would be written as: Ax Ay [P(x) ^ Q(y)] v R(f(x,y)) But from eq. 1, z is in no way related to x and y. So why do we replace z by f(x,y)? Albert pointed out that since disjunction is commutative, we could write Eq. 1 as Ez.R(z) v Ax Ay [P(x) ^ Q(y)] and this, on skolemization, would cause z to be replaced by a constant. I understand that in both the cases we are effectively replacing z by a constant. But, I'm not sure I understand how these two cases are eqivalent. Regards, Rahul. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Mar 23 19:38:20 2005 for ; Wed, 23 Mar 2005 19:38:20 -0500 (EST) for ; Wed, 23 Mar 2005 19:38:19 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 23 Mar 2005 19:30:10 -0500 00:30:10 -0000 j2O0TTBt029652; Wed, 23 Mar 2005 19:29:29 -0500 (EST) -0500 (EST) Date: Wed, 23 Mar 2005 19:29:29 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: Skolemization Comments: To: krishna2@BUFFALO.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Rahul writes: | Let us consider the FOL wff: | | Ax Ay [P(x) ^ Q(y)] v Ez.R(z) .....(1) | | Moving the quantifiers to the left, we have | | Ax Ay Ez [P(x) ^ Q(y)] v R(z) | | >From step 2(a), this would be written as: | | Ax Ay [P(x) ^ Q(y)] v R(f(x,y)) | | But from eq. 1, z is in no way related to x and y. So why do we | replace z by f(x,y)? | | Albert pointed out that since disjunction is commutative, we could | write Eq. 1 as | | Ez.R(z) v Ax Ay [P(x) ^ Q(y)] | | and this, on skolemization, would cause z to be replaced by a constant. | | I understand that in both the cases we are effectively replacing z by | a constant. But, I'm not sure I understand how these two cases are | eqivalent. They're not. But they don't have to be. All that's required is that if the original wff is satisfiable, then so is its Skolemized counterpart. Here's one way to think about this: The Skolem function f(x,y) in the first version above certainly appears to be a function of x and y, but here's a function of x and y that doesn't depend on x or y: f(x,y)=0 I.e., no matter what values x and y take, f(x,y) is the constant function that returns 0. A silly function, perhaps, but a legitimate one. So, it's plausible that f(x,y) above is the same as the Skolem constant in the 2nd version. For more details, I suggest taking a look at Chang & Lee's text on automated theorem proving; see the citation at: http://www.cse.buffalo.edu/~rapaport/572/S02/unification.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Mar 24 15:08:20 2005 for ; Thu, 24 Mar 2005 15:08:20 -0500 (EST) for ; Thu, 24 Mar 2005 15:08:20 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 24 Mar 2005 15:07:10 -0500 20:07:10 -0000 j2OK7Aso012953 for ; Thu, 24 Mar 2005 15:07:10 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Thu, 24 Mar 2005 15:07:10 -0500 (EST) Date: Thu, 24 Mar 2005 15:07:10 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: UBCSEGSA Grad Conference To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: UBCSEGSA Grad Conference ------------------------------------------------------------------------ Information about the UB CSEGSA Grad Conference, including the call for papers (due today!) is at: http://www.cse.buffalo.edu/~rapaport/563S05/flyer2.pdf http://www.cse.buffalo.edu/~rapaport/563S05/Updated_CallForPapers.pdf From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 27 07:11:49 2005 for ; Sun, 27 Mar 2005 07:11:49 -0500 (EST) for ; Sun, 27 Mar 2005 07:11:49 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Mar 2005 07:11:13 -0500 12:11:13 -0000 for ; Sun, 27 Mar 2005 07:11:13 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 27 Mar 2005 07:11:12 -0500 (EST) Date: Sun, 27 Mar 2005 07:11:12 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563/HW 8 QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563/HW 8 QUESTION ------------------------------------------------------------------------ A student writes: | On the asst of hw 8: | | "For partial credit, show all your work, and justify each step using the | algorithm's numbering scheme." | | I don't understand why this is "partial" credit. What then is "full | credit"? could this be "extra" credit? Correct answer with no work = full credit Incorrect answer with work = partial credit Incorrect answer with no work = minimum credit No answer = no credit From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Mar 27 15:04:55 2005 for ; Sun, 27 Mar 2005 15:04:54 -0500 (EST) for ; Sun, 27 Mar 2005 15:04:54 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 27 Mar 2005 15:04:41 -0500 20:04:38 -0000 for ; Sun, 27 Mar 2005 15:04:38 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sun, 27 Mar 2005 15:04:38 -0500 (EST) Date: Sun, 27 Mar 2005 15:04:38 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CF QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 5491 A student writes: | Given a formula (P(x) ^ Q(x)) ( with all the normal interpretations = | of ^, P(x), Q(x) ) | It is fine to rename the variables since there is no quantifier to bound = | either one correct. | So we could safely write this as (P(x) ^ Q(x0)) without losing anything, = | right? Right. You have 2 clauses here, and essentially they are independent. The full wff would have been: Ax[P(x) ^ Q(x)] which is logically equivalent to: Ax.P(x) ^ AxQ(x) which is logically equiv to: Ax.P(x) ^ Ax0.Q(x0) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 28 11:20:04 2005 for ; Mon, 28 Mar 2005 11:20:03 -0500 (EST) for ; Mon, 28 Mar 2005 11:20:03 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Mar 2005 11:19:31 -0500 16:19:31 -0000 j2SGJVso029608 for ; Mon, 28 Mar 2005 11:19:31 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Mar 2005 11:19:31 -0500 (EST) Date: Mon, 28 Mar 2005 11:19:31 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUANTIFIER RULES OF INFERENCE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: QUANTIFIER RULES OF INFERENCE ------------------------------------------------------------------------ I have put on our website a chapter from my logic text that discusses the introduction and elimination rules for the universal and existential quantifiers, in all their gory detail :-) Go to: http://www.cse.buffalo.edu/~rapaport/563S05/fol.html#qfrrules or directly to: http://www.cse.buffalo.edu/~rapaport/563S05/schagrinetal85-qfrrules.pdf From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 28 14:06:29 2005 for ; Mon, 28 Mar 2005 14:06:28 -0500 (EST) for ; Mon, 28 Mar 2005 14:06:28 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Mar 2005 14:05:14 -0500 19:05:14 -0000 j2SJ5Eso001981 for ; Mon, 28 Mar 2005 14:05:14 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Mar 2005 14:05:14 -0500 (EST) Date: Mon, 28 Mar 2005 14:05:14 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 7 ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 7 ANSWERS ------------------------------------------------------------------------ p. 28, #1: ---------- Answers will vary! i) Here are two possible answers: Answer 1: Let D = set of humans [[mary]] = Mary (a female) [[Father(x,y)]] = [[x]] is the father of [[y]] Makes (a) false, since "is the father of" is not transitive (b) true, because the antecedent is always false (c) true, because antecedent always false Answer 2: Let D = {1,2,4,9} [[a]] = [[b]] = 9 [[Double(x,y)]] = "[[x]] = 2*[[y]]" (NB: The quotes are there to remind you that the thing on the RHS of the first "=" sign is a wff of FOL. Note that the first occurrence of "=" is in the *metalanguage*, while the second occurrence (within the quotes) is in FOL itself (the "object" language)) This also makes (a) false, (b) & (c) true. ii) Answer 1: Let D = set of all humans [[mary]] = Mary (a female) [[Brothers(x,y)]] = [[x]] and [[y]] are brothers Makes (a) true, because "are brothers" is transitive (ignoring the technical notion of a "half brother") (b) false; you can't be your own brother (c) true, because antecedent always false Answer 2: Let D = {brother1, brother2, brother3 : all 3 are each other's brothers} [[a]] = brother1 [[b]] = brother2 [[Brother(x,y)]] = [[x]] is a brother of [[y]] (Note that this is *not* the same predicate as "Brothers" above!) iii) Let D = N = {n : n is a natural number} [[a]] = [[b]] = 0 [[LEQ(x,y)]] = "[[x]] <= [[y]]" (i.e., [[x]] is less than or equal to [[y]]) Makes (a) true, because <= is transitive (b) true, because of def of <= (c) false, because antecedent true for all y (0 <= y), but consequent false for all x != 0. Grading: For each problem: D: 0,1,2,3 interpretations of preds/consts: 0,3,6,9 explanations of a's truth value: 0,1,2,3 explanations of b's truth value: 0,1,2,3 explanations of c's truth value: 0,1,2,3 Subtotal = 21 points per problem, Total = 63 points ======================================================================== p. 29, #3: ---------- a) Let "A" and "E" be the universal and existential quantifiers, let "==" be the biconditional, let ">" be the material conditional, and let "-" be negation. For clarity, I omit unnecessary parentheses. (i) -Ex.Elt(x,x) (ii) AxAy[Sub(x,y) == (Az[Elt(z,x) > Elt(z,y)])] (iii) AxAyAz[Elt(z, u(x,y)) == (Elt(z,x) v Elt(z,y))] Grading: 0,1,2,4 points each; subtotal = 12 points ------------------------------------------------------------------------ b) Here, "I" is the book's capital script I. Show that that T |= AxAy.Sub(x, u(x,y)) (where T = the 3 wffs above) I.e., show that, if (1), then (2), where: (1) I,mu |= T (2) I,mu |= AxAy.Sub(x, u(x,y)) It will be easier to do this if we ensure that all variables are unique; this is easier to do with subscripts, but they get messy in Courier font like this, so I'll just range through lower-case letters, *NONE* of which, in what follows, are constants!!: So, show that, if (1) I,mu |= {-Ea.Elt(a,a), AbAc[Sub(b,c) == (Ad[Elt(d,b) > Elt(d,c)])], AeAfAg[Elt(g, u(e,f)) == (Elt(g,e) v Elt(g,f))] } then (2) I,mu |= AhAi.Sub(h, u(h,i)) To do this, assume (1) and show (2). More notation: Let mu-x' represent a variable-assignment that differs from mu at most on x, for any variable "x". >From (1), we have: For all mu-a',...mu-g': I,mu-a',...mu-g'|= {-Elt(a,a), (3) Sub(b,c) == (Ad[Elt(d,b) > Elt(d,c)]), (4) Elt(g, u(e,f)) == (Elt(g,e) v Elt(g,f))} (5) We need to show: For all mu-h',mu-i': I,mu-h',mu-i'|= Sub(h, u(h,i)) >From (4), we have: For all those mu's: I,mu-b',mu-c'|=Sub(b,c) iff I,mu-d'|=(Elt(d,b) > Elt(d,c)) So, all we have to do is show: For all mu-d',mu-h',mu-i': I,mu-d',mu-h',mu-i'|= (Elt(d,h) > Elt(d,u(h,i))) Now I'm going to get a bit sloppy in the interests of clarity: To show the above, suppose that, for all those mu's: |=Elt(d,h) and show |=Elt(d,u(h,i)) (where, modulo my sloppiness, both double-turnstiles have an appropriate sequence of I's and mu's in front of them). In view of (5), to show |=Elt(d,u(h,i)), we just have to show |= (Elt(d,h) v Elt(d,i)) But we have just assumed |=Elt(d,h) So (by an application of "or"-Intro in the metalogic!), we have |= (Elt(d,h) v Elt(d,i)) and, modulo some bookkeeping, we're done! Grading: 0,3,6,9 Total: 21 points. ------------------------------------------------------------------------- p. 30, #4: ---------- [[barber]] = the barber [[Shaves(x,y)]] = [[x]] shaves [[y]] a) Ax[-Shaves(x,x) > Shaves(barber, x)] b) Ax[Shaves(barber, x) > -Shaves(x,x)] Show: if I |= (a) then not(I |= (b)): Suppose I |= (a) Then I |= -Shaves(barber, barber) > Shaves(barber, barber) So, I |= Shaves(barber, barber) v Shaves(barber, barber) So, I |= Shaves(barber, barber) So, not(I |= -Shaves(barber, barber)) So, not(I |= -Shaves(barber, barber) v -Shaves(barber, barber)) So, not(I |= Shaves(barber, barber) > -Shaves(barber, barber)) So, not(I |= Ax[Shaves(barber, x > -Shaves(x,x)])) Grading: syntax & semantics: 0,1,2,3 representation (a) 0,1,2,3 representation (b) 0,1,2,3 proof: 0,3,6,12 Total = 21 ======================================================================== Grand total = 63 + 21 + 21 = 105 463 both 563 A 99-105 A- 94-99 B+ 89-93 B 83-88 B- 77-82 C+ 71-76 C 59-70 36-70 C- 48-58 D+ 36-47 D 19-35 F 0-18 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Mar 28 14:27:27 2005 for ; Mon, 28 Mar 2005 14:27:27 -0500 (EST) for ; Mon, 28 Mar 2005 14:27:27 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 28 Mar 2005 14:26:55 -0500 19:26:07 -0000 j2SJQ7so003027 for ; Mon, 28 Mar 2005 14:26:07 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Mon, 28 Mar 2005 14:26:07 -0500 (EST) Date: Mon, 28 Mar 2005 14:26:07 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PROJECT 1 UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROJECT 1 UPDATE ------------------------------------------------------------------------ As announced in lecture today, I am hereby extending the due date for Project 1 to: Friday, April 15 (and I will modify the project webpages to reflect this) In class, I also suggested the following strategies for handling this project. These are suggestions only! If you are well on the way to completing the project without taking these suggestions into account, keep on with what you're doing! 1. Treat the project as consisting of separable modules: a) the clause-form algorithm * this one you should be able to do now b) the unification algorithm * you can hold off on writing this until we finish covering it in class (unless you're reading ahead in the text :-) c) a resolution applier that takes 2 clauses as input and that outputs their "resolvent", e.g., for the propositional case: Res([phi, alpha-bar], [-phi, beta-bar]) = [alpha-bar, beta-bar] (where "phi" is a propositional wff in CF, and "alpha-bar" and "beta-bar" are disjunctions of wffs in CF) This is a function that you should be able to implement now! For the FOL case, we have to have something slightly more complex, since we might need to unify the phi's before resolving them. Let's suppose that UNIF is our unification algorithm (which we have not yet covered in lecture). Then the resolution function for FOL looks like this: Res([UNIF(phi1), alpha-bar], [-UNIV(phi2), beta-bar]) = [alpha-bar, beta-bar] (where "phi1" and "phi2" are wffs in CF that need to be unified before being able to resolve them; when they are unified, UNIF(phi1) will be identical to UNIF(phi2), so in the 2 input clauses to Res, we'll have a wff and its negation, just as we do in the propositional case). Note, therefore, that this version of Res includes the propositional case, since, in the propositional case, phi1=phi2 to begin with and since UNIF(phi), where phi is a propositional wff is just phi itself (there are no variables to substitute for). This means that you can, right now, write your version of Res, using a "stub" procedure for UNIF. In the propositional case, it will just return its argument (i.e., it will be a no-op), and in the FOL case, for now, it can return an error message ("Sorry, this is not yet implemented"). Then, when you feel comfortable about writing UNIF, you can just plug it in. 2. There is also the issue of combining all of these into a single program that takes as input an argument S |- alpha and uses refutation, resolution, and unification (applied to the CF forms of S and -alpha) to determine, by deriving [], whether alpha follows from S. You can leave that for last. The CF algorithm can stand on its own. The Res algorithm can either take its input pairs from the user, or can offer the user a choice of wffs from CF(S) union CF(-alpha) (unioned with any new wffs derived via resolution of previous wffs), ending when [] is added to the set. 3. There are also the individual exercises that, in the project specs, I have asked you to do at least by hand whether or not you are able to complete the implementation. Most of those can be done now (the unification examples can wait till we cover it in class). 4. Finally, there's the report. Don't wait till the last minute on this! There's no reason you couldn't start writing the report now, describing what each module does and how they fit together, inserting the actual code and sample runs as they become available. Even if you fail to complete any of the implementations, you can still have a report! (Of course, in that case, it should have a section describing the bugs in whatever code you do have, and how you would have tried eliminating those bugs had I given you yet another extension :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 29 13:54:16 2005 for ; Tue, 29 Mar 2005 13:54:15 -0500 (EST) for ; Tue, 29 Mar 2005 13:54:15 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 29 Mar 2005 13:53:44 -0500 18:53:44 -0000 j2TIriso010324 for ; Tue, 29 Mar 2005 13:53:44 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 29 Mar 2005 13:53:44 -0500 (EST) Date: Tue, 29 Mar 2005 13:53:44 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUESTION ABOUT HW #8, problem 3 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A student writes: | Were we given any rules for handling equality? There's nothing that I | saw that makes it special... Can we just represent it as a predicate | with an odd form when we generate clauses? For our purposes, no special rules for "=" are needed. Treat it as a 2-place predicate no different from any other (except for its use of the more familiar infix notation). However, Sect. 4.3.7 does give some discussion to special considerations for equality, with a special rule of inference called Paramodulation. We will not cover that topic this semester. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Mar 29 14:30:14 2005 for ; Tue, 29 Mar 2005 14:30:14 -0500 (EST) for ; Tue, 29 Mar 2005 14:30:13 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 29 Mar 2005 14:29:29 -0500 19:29:29 -0000 j2TJTSso010510 for ; Tue, 29 Mar 2005 14:29:28 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Tue, 29 Mar 2005 14:29:28 -0500 (EST) Date: Tue, 29 Mar 2005 14:29:28 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: SIMULTANEOUS SUBSTITUTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: SIMULTANEOUS SUBSTITUTION ------------------------------------------------------------------------ Simultaneous substitution is discussed in more depth in: Gries, David (1981), The Science of Programming (New York: Springer-Verlag), Sect. 4.4 ("Textual Substitution"), pp. 79-82. Here is an example from that text, converted to our notation: Let t = {x/f(x,y), y/z} be a substitution Let F = g(x,x,y) be an expression Then Ft = g(f(x,y), f(x,y), z) Quoting Gries, this example "illustrates the fact that the substitution must be simultaneous; if one first replaces all occurrences of x and *then* replaces all occurrences of y, the result is g(f(x,z), f(x,z), z), which is not the same. In general," E{x/u, y/v} "can be different from" E{x/u}{y/v}. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Apr 2 22:11:51 2005 for ; Sat, 2 Apr 2005 22:11:50 -0500 (EST) for ; Sat, 2 Apr 2005 22:11:50 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 2 Apr 2005 22:11:14 -0500 03:11:14 -0000 for ; Sat, 2 Apr 2005 22:11:14 -0500 (EST) cse563-sp05-list@listserv.buffalo.edu; Sat, 2 Apr 2005 22:11:14 -0500 (EST) Date: Sat, 2 Apr 2005 22:11:14 -0500 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: RESOLUTION: TYING UP LOOSE ENDS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 5862 ------------------------------------------------------------------------ Subject: CSE 4/563: RESOLUTION: TYING UP LOOSE ENDS ------------------------------------------------------------------------ One thing I didn't get a chance to do at the end of class on Friday was to take you on a tour of the "Automated Theorem Proving" website, so highlight some features of it, so let me do that now. At http://www.cse.buffalo.edu/~rapaport/563S05/fol-atp.html you will find the following items: # Resolution Proof in Natural-Deduction Format This is the propositional resolution proof that we did in lecture, rewritten in the style of our natural-deduction system. # Notes on the Unification Algorithm [PDF] This contains the notes from lecture with the definitions of substitution, unifier, etc. # Unification Algorithm (PDF file) This is the algorithm that I handed out. # Unification Example (PDF file) This is the example that I handed out. # Further Thoughts on Unification This just has a couple of comments on unification including an exercise "left to the reader", and some references. # Another Resolution Example This is, as it says, another example. # Examples of Resolution with Answer Literals: These are 2 more examples, this time showing how you can conjoin an "Answer" predicate to the query in such a way that, when resolution is done, instead of getting the empty clause, you get a clause containing the answer to your question! You should browse through these pages, study the examples, and maybe, if you have time now or after the course is over, take a look at some of the references. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 3 00:46:41 2005 for ; Sun, 3 Apr 2005 00:46:41 -0500 (EST) for ; Sun, 3 Apr 2005 00:46:41 -0500 (EST) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 3 Apr 2005 00:46:20 -0500 05:46:20 -0000 j335kKn9002863 for ; Sun, 3 Apr 2005 00:46:20 -0500 (EST) 00:46:20 -0500 (EST) Date: Sun, 3 Apr 2005 00:46:19 -0500 Reply-To: Albert Goldfain From: Albert Goldfain Subject: HW#7 problem 1 (B & L p.28) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: Junk X-UID: 5874 I want to use most of next recitation for Unification and Resolution practice, so I will make some comments here on your interpretations for HW #7 problem 1 (this was the one where you had to give a semantic interpretation that made two wffs true and one false) First of all, remember that a semantic interpretation for FOL MUST include a specification of the domain of discourse, an interpretation of the predicates and function symbols (there were no function symbols in the exercise however) and an interpretation of the constants. It CANNOT be just an assignment of T/F values...since [[P(x,y)]] = T is meaningless in FOL (T/F assignments are only meaningful in propositional logic). Also, remember that your constants must be constant! That is, they cannot vary over your domain. Thus, if your domain D = the set of all people, then you CANNOT let [[a]] = any woman (this restricts a in the domain, but it is still a variable). You must pick out a particular woman (i.e., a constant). This time around I did not take off for this...but please make sure your constants are constant from now on. Dr. Rapaport asked me to "record any interestingly different answers, for future reference" for this problem. Most of the 100% correct answers were not so interesting (or, not so interestingly different from Dr. Rapaport's answers :) The interesting answers were the ones in which the domain was wide enough to allow counterexamples to some of your claims for truth. Apparently, some of you lead "sheltered" enough lives to think that our world has no: twins (or triplets, quadruplets, ...), polygamy, or androgyny/hermaphrodites :) [Look any of these up in the dictionary if needed]. If, as a "knowledge engineer", you claim your "system" will work for the "domain of all people" then you have to consider all of the "special cases" (Especially since one engineer's "special" might be another engineer's "normal"). Usually the answer is to restrict the domain in some way (e.g., "all people except those in polygamous marriages"). This will show that you know the limits of your "system"...which is good for any kind of engineering! But this is just a homework...and since you weren't asked to develop a real-world capable agent, I was fairly kind to the "sheltered" souls :) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 3 12:49:45 2005 for ; Sun, 3 Apr 2005 12:49:44 -0400 (EDT) for ; Sun, 3 Apr 2005 12:49:44 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 3 Apr 2005 12:48:35 -0400 16:48:35 -0000 for ; Sun, 3 Apr 2005 12:48:35 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Sun, 3 Apr 2005 12:48:34 -0400 (EDT) Date: Sun, 3 Apr 2005 12:48:34 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MAJOR SYLLABUS UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: MAJOR SYLLABUS UPDATE ------------------------------------------------------------------------ I have revised the syllabus to indicate more clearly what we have accomplished so far and what I plan to do during the rest of the semester. (This is *not* likely to be the last update, however :-) For the latest version, go to: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html#dates From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 4 09:37:00 2005 for ; Mon, 4 Apr 2005 09:36:59 -0400 (EDT) for ; Mon, 4 Apr 2005 09:36:59 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 4 Apr 2005 09:32:39 -0400 13:32:39 -0000 j34DWdso005256 for ; Mon, 4 Apr 2005 09:32:39 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Mon, 4 Apr 2005 09:32:38 -0400 (EDT) Date: Mon, 4 Apr 2005 09:32:38 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PREPARING REPORT IN LATEX To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A student writes: | I've written my program and documented the code. I'm to the point where the | paper is all that's left. | | Could you post a listing of useful latex symbols to the listserv? Well, depends on what you think might be useful. The standard source of such info is the LaTeX manual by Lamport. I don't know if there's an online list of the symbols. | One such | example is: | $\circ$ for composition | | The $ character is to enter and escape math mode, as I'm sure you're well | aware. You might want to state that to beginners at LaTeX, and perhaps even | provide the students with a site where they could learn how to use LaTeX. Don't know of any; I learned it from the manual. I can, however, prepare a LaTeX template that you can use. Give me a day or so. | Some symbols I'd like are: | conjunction $\wedge$ | disjunction $\vee$ | implication $\supset$ | biconditional (or congruency) $\equiv$ | negation $\neg$ | Is sig-alt-release an acceptable document style? Or is there another one | that you prefer? I'm not familiar with that style. The standard invocation is: \documentclass[11pt]{article} \usepackage{amssymb,fullpage,times,mathptm} The "article" class is the standard one for a report such as this. The "amssymb" package gives you a wider variety of symbols, though you probably won't need them for your report. The "fullpage" package sets correct margins (1" on each side) The "times" package uses the Times Roman font instead of the standard LaTeX "computer modern" font. The main advantage besides looking nicer is that it creates better PDF files. The "mathptm" font puts symbols into Times Roman. Let me know if any of you have other LaTeX questions. From stock@cse.Buffalo.EDU Mon Apr 4 09:51:40 2005 for ; Mon, 4 Apr 2005 09:51:40 -0400 (EDT) for ; Mon, 4 Apr 2005 09:51:40 -0400 (EDT) In-Reply-To: <200504041332.j34DWcIx005255@wasat.cse.buffalo.edu> Cc: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU From: Matthew Stock Subject: Re: CSE 4/563: PREPARING REPORT IN LATEX Date: Mon, 4 Apr 2005 09:51:39 -0400 To: "William J. Rapaport" On Apr 4, 2005, at 9:32 AM, William J. Rapaport wrote: > | Could you post a listing of useful latex symbols to the listserv? > I like using http://omega.albany.edu:8008/Symbols.html as a reference. You can also use a latex editor that knows about symbols, such as iTeXMac (for the Mac). -Matt From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 4 11:56:19 2005 for ; Mon, 4 Apr 2005 11:56:19 -0400 (EDT) for ; Mon, 4 Apr 2005 11:56:19 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 4 Apr 2005 11:50:04 -0400 15:50:03 -0000 j34Fo3so007850 for ; Mon, 4 Apr 2005 11:50:03 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Mon, 4 Apr 2005 11:50:03 -0400 (EDT) Date: Mon, 4 Apr 2005 11:50:03 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: LATEX TEMPLATE & FORMATTING YOUR REPORT To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: LATEX TEMPLATE & FORMATTING YOUR REPORT ------------------------------------------------------------------------ For those of you using LaTeX for your report (and you should all be using LaTeX for your report :-), I have put a template that you can use on the website at: http://www.cse.buffalo.edu/563S05/template.tex ***** Whether or not you use LaTeX *****: Please be sure to read the syllabus for instructions on how to prepare your report: http://www.cse.buffalo.edu/~rapaport/563S05/syl.html#projectpolicies AND Please be sure to read my online document "How to Write" for further general instructions that *must* be followed! http://www.cse.buffalo.edu/~rapaport/howtowrite.html From rapaport@cse.Buffalo.EDU Mon Apr 4 12:45:16 2005 Mon, 4 Apr 2005 12:45:16 -0400 (EDT) Date: Mon, 04 Apr 2005 12:45:15 -0400 From: "William J. Rapaport" Organization: SUNY Buffalo Computer Science & Engineering User-Agent: Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.0.1) Gecko/20020920 Netscape/7.0 To: snerg@cse.Buffalo.EDU, cse563-sp05-list@listserv.buffalo.edu Subject: CSE 4/563: Semantic Networks I just discovered this very nice webpage on semantic nets, which, among other things, discusses SNePS! -- ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html From rapaport@cse.Buffalo.EDU Tue Apr 5 14:26:40 2005 Tue, 5 Apr 2005 14:26:39 -0400 (EDT) for ; Tue, 5 Apr 2005 14:26:39 -0400 (EDT) Tue, 5 Apr 2005 14:26:38 -0400 (EDT) Tue, 5 Apr 2005 14:26:38 -0400 (EDT) Date: Tue, 5 Apr 2005 14:26:38 -0400 (EDT) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu, snerg@cse.Buffalo.EDU Subject: A paying job in ontology, with social implications | Date: Tue, 5 Apr 2005 13:12:36 -0400 | From: "Smith, Barry" | Subject: The Origins of Property in Developing Economies | | Some of you will remember the conference on "The Mystery of Capital and the | Construction of Social Reality" (featuring Hernando de Soto and John | Searle) held in Buffalo in 2003. | | De Soto is the head of the Institute for Liberty and Democracy | | http://www.ild.org.pe/tmoc/cp1-en.htm | | in Lima, Peru, described by Bill Clinton as 'the most promising | anti-poverty initiative in the world'. | | The ILD's work consists in finding ways to formalize the currently informal | (extra-legal) property rights enjoyed by many poor people in developing | countries. As a result of the Buffalo conference, de Soto has now put me in | charge of a project, funded by the US Agency for International Development, | that is designed to establish the philosophical underpinnings of this work, | effectively by addressing broadly ontological questions such as: what is a | property right? what is a formal (legal) property right? what are the | origins of property rights? what are the legal, social, economic, cultural, | cognitive preconditions of property rights? Etc.) | | Under the auspices of this project the ILD has already hired Gloria Zuniga, | a graduate of the Department, to the position of Director of Research in | Lima. We will also need other kinds of assistance, and we will have | resources to pay interested UB graduate students to carry out specific | kinds of research work in the next months and years. If anyone thinks they | might be interested, please contact me as soon as possible. | | Barry Smith | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Apr 5 15:43:10 2005 for ; Tue, 5 Apr 2005 15:43:10 -0400 (EDT) for ; Tue, 5 Apr 2005 15:43:09 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 5 Apr 2005 14:26:39 -0400 18:26:39 -0000 j35IQcso013766; Tue, 5 Apr 2005 14:26:38 -0400 (EDT) 14:26:38 -0400 (EDT) Date: Tue, 5 Apr 2005 14:26:38 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: A paying job in ontology, with social implications Comments: To: snerg@cse.Buffalo.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | Date: Tue, 5 Apr 2005 13:12:36 -0400 | From: "Smith, Barry" | Subject: The Origins of Property in Developing Economies | | Some of you will remember the conference on "The Mystery of Capital and the | Construction of Social Reality" (featuring Hernando de Soto and John | Searle) held in Buffalo in 2003. | | De Soto is the head of the Institute for Liberty and Democracy | | http://www.ild.org.pe/tmoc/cp1-en.htm | | in Lima, Peru, described by Bill Clinton as 'the most promising | anti-poverty initiative in the world'. | | The ILD's work consists in finding ways to formalize the currently informal | (extra-legal) property rights enjoyed by many poor people in developing | countries. As a result of the Buffalo conference, de Soto has now put me in | charge of a project, funded by the US Agency for International Development, | that is designed to establish the philosophical underpinnings of this work, | effectively by addressing broadly ontological questions such as: what is a | property right? what is a formal (legal) property right? what are the | origins of property rights? what are the legal, social, economic, cultural, | cognitive preconditions of property rights? Etc.) | | Under the auspices of this project the ILD has already hired Gloria Zuniga, | a graduate of the Department, to the position of Director of Research in | Lima. We will also need other kinds of assistance, and we will have | resources to pay interested UB graduate students to carry out specific | kinds of research work in the next months and years. If anyone thinks they | might be interested, please contact me as soon as possible. | | Barry Smith | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 6 09:36:42 2005 for ; Wed, 6 Apr 2005 09:36:42 -0400 (EDT) for ; Wed, 6 Apr 2005 09:36:41 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 6 Apr 2005 09:36:22 -0400 13:36:22 -0000 j36DaLso017633 for ; Wed, 6 Apr 2005 09:36:21 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Wed, 6 Apr 2005 09:36:21 -0400 (EDT) Date: Wed, 6 Apr 2005 09:36:21 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: TENTATIVE GRADING SCHEME FOR PROJECT 1 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: TENTATIVE GRADING SCHEME FOR PROJECT 1 ------------------------------------------------------------------------ This document is intended to help you prepare your report on Project 1, by giving you some idea of the kinds of things we will be looking for. It is subject to change, but probably not in any significant way. (If we change it, we'll let you know.) ======================================================================== Report + Sample Runs: --------------------- (Note: Please put your sample runs in the body of your report, rather than in the appendix with the code. It might help if you think of the *audience* for your report as someone you are tutoring in logic for AI.) POINTS (*) (*) Point possibilities are of the form "0 n 2n 3n", where 0 means that the work is missing, n points is for an unsatisfactory response, 3n points is for a satisfactory response, and 2n points is for partial credit. ------------------------------------------------------------------------- (1) (a) Description of Clause Form and of your algorithm 0 5 10 15 (b) Correct answers to (b) (by hand) 0 5 10 15 Annotated sample run for (b) 0 5 10 15 Total = 45 points ------------------------------------------------------------------------- (2) (a) Description of unification and of your algorithm 0 5 10 15 (b) Correct answers to (i)-(v) (by hand) 0 1 2 3 for each; total = 15 Annotated sample runs for (i)-(v) 0 1 2 3 for each; total = 15 Total = 45 points ------------------------------------------------------------------------- (3) (a) Description of automated theorem proving using resolution, unification, and refutation 0 5 10 15 (b) Solution (by hand) of the example 0 5 10 15 Annotated sample run for (b) 0 5 10 15 Total = 45 points Grand subtotal = 135 points ------------------------------------------------------------------------- Appendix Containing Documented Code: ----------------------------------- clause form: code: 0 5 10 15 documentation: 0 5 10 15 unification algorithm: code: 0 5 10 15 documentation: 0 5 10 15 resolution: code: 0 5 10 15 documentation: 0 5 10 15 Grand subtotal = 90 points ------------------------------------------------------------------------- Optional bug report: 0 5 10 15 Let's assume that there are parts of your project that you were unable to get to work. If you write an analysis of the problems you encountered and suggest some debugging strategies that you would have tried had I given you another couple of days to work on it, you can earn up to 15 points extra credit. ------------------------------------------------------------------------- Grand total points = 225 472 both 572 --- ---- --- A 214-225 A- 201-213 B+ 189-200 B 176-188 B- 164-175 C+ 151-163 C 126-150 76-150 C- 101-125 D+ 76-100 D 39-75 F 0-38 From stock@cse.Buffalo.EDU Wed Apr 6 14:22:04 2005 for ; Wed, 6 Apr 2005 14:22:04 -0400 (EDT) for ; Wed, 6 Apr 2005 14:22:04 -0400 (EDT) In-Reply-To: <200504061336.j36DaLgB017632@wasat.cse.buffalo.edu> Cc: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU From: Matthew Stock Subject: Re: CSE 4/563: TENTATIVE GRADING SCHEME FOR PROJECT 1 Date: Wed, 6 Apr 2005 14:22:03 -0400 To: "William J. Rapaport" On Apr 6, 2005, at 9:36 AM, William J. Rapaport wrote: > (1) (a) Description of Clause Form and of your algorithm 0 5 10 > 15 > > (b) Correct answers to (b) (by hand) 0 5 10 > 15 > > Annotated sample run for (b) 0 5 10 > 15 > > Total = 45 > points Does this breakdown mean that we must provide a "hand generated" answer (presumably using the algorithms provided), even if our code is operational, to get full credit? Thanks, -Matt From rapaport@cse.Buffalo.EDU Wed Apr 6 16:11:18 2005 Wed, 6 Apr 2005 16:11:15 -0400 (EDT) Wed, 6 Apr 2005 16:11:15 -0400 (EDT) Wed, 6 Apr 2005 16:11:15 -0400 (EDT) Wed, 6 Apr 2005 16:11:15 -0400 (EDT) Date: Wed, 6 Apr 2005 16:11:15 -0400 (EDT) From: "William J. Rapaport" To: rapaport@cse.Buffalo.EDU, stock@cse.Buffalo.EDU Subject: Re: CSE 4/563: TENTATIVE GRADING SCHEME FOR PROJECT 1 Cc: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | From stock@cse.Buffalo.EDU Wed Apr 6 14:22:04 2005 | ... | Does this breakdown mean that we must provide a "hand generated" answer | (presumably using the algorithms provided), even if our code is | operational, to get full credit? Yes. I quote from the project specifications: "NOTE: Please do all exercises at least by hand (in addition to, or instead of, implementing them in a programming language) as part of your report." The idea is to show an isomorphism between the hand-generated "gold standard" and your automated solution. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 6 17:32:50 2005 for ; Wed, 6 Apr 2005 17:32:50 -0400 (EDT) for ; Wed, 6 Apr 2005 17:32:50 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 6 Apr 2005 17:26:51 -0400 21:26:51 -0000 j36LQpn9007078 for ; Wed, 6 Apr 2005 17:26:51 -0400 (EDT) 17:26:51 -0400 (EDT) Date: Wed, 6 Apr 2005 17:26:51 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Resolution from this week's recitation To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, As promised, I wrote up a "neat" version of the "Curiosity killed the cat" resolution we did in recitation. You can find the PDF at: http://www.cse.buffalo.edu/~ag33/ResolutionExample.pdf I personally don't know how to do large resolution trees in LaTex so I gave the resolution step as a sequence of calls to a "resolve" function with clause pairs. Even if I *did* know how to draw these trees in LaTex, I don't know if the result would be any cleaner than the complicated one I drew on the board :-) You should be able to trace it though though. Albert ps...the resolution we did in the monday and wednesday recitation was not identical in the order of resolved clauses...the posted solution is *one* of multiple resolution proofs. pps...if there are any problems with the proof then let me know ASAP From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 6 22:27:43 2005 for ; Wed, 6 Apr 2005 22:27:43 -0400 (EDT) for ; Wed, 6 Apr 2005 22:27:42 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 6 Apr 2005 22:24:00 -0400 02:23:59 -0000 j372Nxn9021676 for ; Wed, 6 Apr 2005 22:23:59 -0400 (EDT) 22:23:59 -0400 (EDT) Date: Wed, 6 Apr 2005 22:23:59 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: A nice article on automated theorem proving To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU The Economist has an article on automated theorem provers and their place in the history of "proof": http://www.economist.com/science/displayStory.cfm?story_id=3809661 In case you needed motivation finishing up Project #1 :) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 7 19:02:29 2005 for ; Thu, 7 Apr 2005 19:02:29 -0400 (EDT) for ; Thu, 7 Apr 2005 19:02:28 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 7 Apr 2005 19:02:06 -0400 23:02:06 -0000 for ; Thu, 7 Apr 2005 19:02:06 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 7 Apr 2005 19:02:06 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 7 Apr 2005 19:02:06 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MGU QUESTION To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% A student writes: | I was wondering if this is valid when performing unification, having an | mgu thats looks like this{x/y, y/z, z/y}? Does this work? | | I know these replacements are supposed to be done in parallel so that is | why I believe this would not work. This question is hard to answer, because it's making a presupposition, namely that a given substitution (i.e., a set of bindings) either is or is not an MGU. There is really no such thing as "an MGU". Rather, there can be an MGU *for* some wffs. So, show me the wffs that you think this is an MGU for, and I can give you a better answer. But I doubt there are any such wffs, since the particular substitution above is suspect anyway. Which part of the unification algorithm might have generated it? The only way would have been as the result of composing two substitutions. Let's ignore the 3rd element of the above set for a moment, and just focus on {x/y, y/z}. This could not be the result of composing any two substitutions. At best, someone might think it's {x/y}o{y/z}, but by the definition of composition, {x/y}o{y/z} = {x/y{y/z}, y/z} = {x/z, y/z}, which is not our set above. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 7 21:46:21 2005 for ; Thu, 7 Apr 2005 21:46:20 -0400 (EDT) for ; Thu, 7 Apr 2005 21:46:20 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 7 Apr 2005 21:45:36 -0400 for ; Thu, 7 Apr 2005 21:45:33 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 7 Apr 2005 21:45:33 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 7 Apr 2005 21:45:33 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: CSE 4/563: Translating from Propositional Logic to English To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% ------------------------------------------------------------------------ Subject: Re: CSE 4/563: Translating from Propositional Logic to English ------------------------------------------------------------------------ Back in February, I wrote: | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Feb 27 21:21:52 2005 | ... | 3. | This is not directly related to any of the above, but here are two | examples of grammatically correct and meaningful, but hard-to-understand, | sentences of English. Can you figure out what they mean? | | Dogs dogs dog dog dogs. | and | Buffalo buffalo buffalo buffalo buffalo. There have been a few follow-up postings; here's the latest: | Date: Thu, 7 Apr 2005 18:04:45 -0400 (EDT) | From: Paula Chesley | To: "William J. Rapaport" | Subject: Buffalo buffalo buffaloing their Buffalo kin | | Hello, | | A bit of time has passed since you asked us what the deal was with the | 'buffalo' sentence. I don't have my copy of Stephen Pinker's __The | Language Instinct__ in, well, Buffalo, so I had to remember what the heck | the Buffalo^5 (or 6 or 7) sentence means: | | (paraphrase of Pinker) | | A buffalo from Buffalo is a Buffalo buffalo, that could `buffalo' | its kin. Hence: `(Those) Buffalo buffalo (that other) Buffalo | buffalo buffalo (will respectively) buffalo (even different) Buffalo | buffalo'. | | In contrast to Pinker, I certainly don't consider this sentence | grammatical. In fact, it | points to the importance of function words, like the copula and | connectives, in our comprehension of language. | | Paula I've already give you a bit of the history (browse the Listserv archive if you've forgotten), but here are the answers. First, consider this: Mice cats chase eat cheese. Most native speakers of English have no problem understanding this: There are 2 kinds of mice, those that cats chase and those that cats don't chase. The former are mice cats chase. They eat cheese. Now, admittedly, this sounds better with "that" after "Mice": Mice that cats chase eat cheese. But that "that" is omittable in English. Now consider that there are two kinds of dogs, those that other dogs dog (i.e., that are followed along closely behind by other dogs) and those that aren't. The former are dogs dogs dog. They also dog dogs. So, dogs dogs dog dog dogs. Finally, we need a noun whose plural is the same as its singular (so that we don't have to add -s at the end) and that is also a verb. There are actually several possibilities. The one I came up with when I was in grad school was "buffalo": The plural of "buffalo" is also "buffalo", and "to buffalo" means to intimidate (among other things). So, buffalo buffalo buffalo buffalo buffalo; i.e., buffalo who are buffaloed by other buffalo also buffalo still other buffalo. Or, as I mentioned in one of the earlier postings: [[is-a-buffalo(x)]] = x is a buffalo [[buffalos(x,y)]] = x buffalos y Ax[(is-a-buffalo(x) ^ Ey[is-a-buffalo(y) ^ buffalos(y,x)]) > Ez[is-a-buffalo(z) ^ buffalos(x,z)]] As for Pinker, his interpretation is actually a bit different as I recall, and it turns out that the buffalo sentence is syntactically ambiguous! His interpretation is that "buffalo" is also an adjective, indicating a buffalo that comes from Buffalo. So, Buffalo buffalo (i.e., the ones who live in Buffalo) buffalo other Buffalo buffalo. Now: Can you think of any other words that you can play this trick with? From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat Apr 9 09:20:49 2005 for ; Sat, 9 Apr 2005 09:20:49 -0400 (EDT) for ; Sat, 9 Apr 2005 09:20:49 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 9 Apr 2005 09:03:34 -0400 13:03:34 -0000 for ; Sat, 9 Apr 2005 09:03:33 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Sat, 9 Apr 2005 09:03:33 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Sat, 9 Apr 2005 09:03:33 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 8 ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% HW 8 Answers & Grading: ------------------------------------------------------------------------ Notation: "X --> Y" for: X can be rewritten as Y ------------------------------------------------------------------------ 1. (P^Q) v (R^S) --> [(P^Q) v R] ^ [(P^Q) v S] --> (PvR) ^ (QvR) ^ (PvS) ^ (QvS) this is CNF --> {[P,R], [Q,R], [P,S], [Q,S]} this is CF ------------------------------------------------------------------------ ------------------------------------------------------------------------ 2. Note: In what follows, I leave blank space to show where the deleted existential quantifier was, but in fact there are no blank spaces. Thus, e.g., the 5th line is really: AyAuAv.P(...) ExAyEzEwAuAvEt.P(x,y,z,w,u,v,t) --> AyEzEwAuAvEt.P(c,y,z,w,u,v,t), where "c" is a Skolem constant --> Ay EwAuAvEt.P(c,y,fz(y),w,u,v,t), where "fz" is a Skolem function --> Ay AuAvEt.P(c,y,fz(y),fw(y),u,v,t), where "fw" is a new Skolem function --> Ay AuAv .P(c,y,fz(y),fw(y),u,v, ft(y,u,v) ) --> P(c, y, fz(y), fv(y), u, v, ft(y, u, v) ) this is CNF --> {[P(c,y,fz(y),fv(y),u,v,ft(y,u,v))]} this is CF ------------------------------------------------------------------------ ------------------------------------------------------------------------ 3. Ex[S(x) ^ Ay[S(y) > y=x]] --> Ex[S(x) ^ Ay[-S(y) v y=x]] --> ExAy[S(x) ^ (-S(y) v y=x)] --> Ay[S(c) ^ (-S(y) v y=c)] where "c" is a Skolem constant --> S(c) ^ (-S(y) v y=c) this is CNF --> {[S(c)], [-S(y),y=c]} this is CF ------------------------------------------------------------------------ ------------------------------------------------------------------------ 4. ExAy[-S(y,y) > S(x,y)] --> ExAy[--S(y,y) v S(x,y)] --> ExAy[S(y,y) v S(x,y)] --> Ay[S(y,y) v S(b,y)], where "b" is a Skolem constant --> S(y,y) v S(b,y) this is CNF --> {[S(y,y),S(b,y)]} this is CF ------------------------------------------------------------------------ ======================================================================== Grading: 0,1,2,3 points each; total = 12 points, where: 0 = not done 1 = incorrect & not enough work shown for us to figure out what went wrong 2 = incorrect, but shows enough intermediate steps for us to figure out what went wrong 3 = correct clause form (or at least correct CNF) 463 both 563 A 12 A- 11 B+ 10 B 9 B- 8 C+ 7 C 6 4-6 C- 5 D+ 4 D 3 F 0-2 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 10 13:14:07 2005 for ; Sun, 10 Apr 2005 13:14:07 -0400 (EDT) for ; Sun, 10 Apr 2005 13:14:07 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 10 Apr 2005 13:13:51 -0400 for ; Sun, 10 Apr 2005 13:13:50 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Sun, 10 Apr 2005 13:13:50 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Sun, 10 Apr 2005 13:13:50 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: QUESTION ABOUT COMPOSITION OF SUBSTITUTIONS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% ------------------------------------------------------------------------ Subject: CSE 4/563: QUESTION ABOUT COMPOSITION OF SUBSTITUTIONS ------------------------------------------------------------------------ | I have a question about the handout "Notes on the Unification Algorithm". | On the second page, there is an ***UPDATE*** box. It requires us to | eliminate "redundant substitutions" by subtracting {x/t sigma:t sigma=x}. | | I do not understand why that is necessary because if theta has x/t, all | the x's are gone, so sigma has no x's in it. There will never be such | kind of redundancy. I am assuming that theta has | been applied before obtaining sigma. Let me try to clarify the definition of the composition of two substitutions. Call the substitutions S and T; let SoT be the composition of S with T (which, recall, is different from ToS). Let S = {x1/s1, ..., xn/sn} Let T = {y1/t1, ..., ym/tm} Here is an algorithm for computing SoT: 1. Apply T to the terms in the bindings that are members of S: i.e., replace each xi/si in S with xi/siT i.e., apply T to any applicable si in S, to form a new binding xi/siT 2. As a result of doing that, some of the resulting bindings might be "redundant" in the sense that siT = xi i.e., the result of applying T to si might be to turn si into xi but such a binding is "redundant" in the sense that binding xi to itself is a no-op. 3. Next, take the union of the thus-modified S with T, but... if this results in two bindings in the union that have the same left side, i.e., if it results in two bindings that inconsistently require you to bind two different terms to the same variable, then delete the binding that originated in T and keep the (possibly modified, from step 1) binding that originated in S If you still have questions, please try to frame them in terms of the above algorithm. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 11 09:22:36 2005 for ; Mon, 11 Apr 2005 09:22:36 -0400 (EDT) for ; Mon, 11 Apr 2005 09:22:35 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 11 Apr 2005 09:21:48 -0400 13:21:48 -0000 for ; Mon, 11 Apr 2005 09:21:47 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Mon, 11 Apr 2005 09:21:47 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Mon, 11 Apr 2005 09:21:47 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: CSE 4/563: QUESTION ABOUT COMPOSITION OF SUBSTITUTIONS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% ------------------------------------------------------------------------ Subject: Re: CSE 4/563: QUESTION ABOUT COMPOSITION OF SUBSTITUTIONS ------------------------------------------------------------------------ Another student replies: | I noticed what the student ... is saying too. There isn't anything wrong | with the algorithm you have supplied, but rather as side affect of our | unification algorithm. We will never get the conditions for rule 3 below. | Since after every substitution, we apply that substitution to the atomic | formulas were are trying to unify. | Ex: | Unify(P(x,f(a)),P(b,x)) //assuming x is a variable, f(a) is a function | symbol, and a is a constant | Here we will have the first substitution being x/b. In which case well then | call unify again, which will look like | Unify(P(b,f(a)),P(b,b)) | Here obviously it will then fail, but all occurrences later of x will be | replaced by the substitution, and thus we could never have a multiple | mapping of x. | So that's why they are saying some situations in the compose algorithm may | not occur. Let me know if this makes more sense, or if I'm wrong :) I think the problem both of you see arises from a subtle curiosity in our version of the unification algorithm. Note that it does the following: 1. It computes an updated unifier by composing the previous unifier with the current occurs-checked substitution (this is in the first step under "begin" in step 3) 2. Next, it computes an updated W by applying what I am calling the "occurs-checked substitution" to the previous W. You are probably right (though I'd like to see an example) that, because of this piecemeal way of doing things, part of the composition definition is not needed. However: A. If the updated W were computed by applying the updated unifier to the *original* W0, then you (probably) *would* still need the full definition of composition. B. In any case, to compute an MGU, you must use the full, correct definition of composition (and why on earth would you want to use an incorrect definition anyway, just because it "works"? :-) That's reminiscent of the old assembly-language tricks that saved a bit of storage here and there by reusing data that just by accident happened to be available somewhere instead of having to "waste" a memory register by storing it in advance; it may be efficient, but it's lousy style). So, bottom line: if you update W "piecemeal", you are probably correct that that extra step of eliminating redundancy in the composed unifier is itself redundant. However, if you update W by going back to W0 and applying the MGU, then--in order to have the correct MGU (which, remember, is the main output of the algorithm!)--you need to use the correct definition of composition. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Apr 12 08:09:00 2005 for ; Tue, 12 Apr 2005 08:09:00 -0400 (EDT) for ; Tue, 12 Apr 2005 08:09:00 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 12 Apr 2005 08:08:09 -0400 for ; Tue, 12 Apr 2005 08:08:08 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Tue, 12 Apr 2005 08:08:08 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Tue, 12 Apr 2005 08:08:08 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CogSci: Len Talmy on language's attention system To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6003 ------------------------------------------------------------------------ Subject: CSE 4/563: CogSci: Len Talmy on language's attention system ------------------------------------------------------------------------ This may or may not be directly or indirectly related to KRR (I strongly suspect, however, that it *is* related), but this may be a once in a lifetime opportunity for some of you to hear Len Talmy speak. Len is one of, if not the, leading figures in cognitive linguistics and he is a wonderful speaker, not so much in terms of his dynamics, but in terms of the frequency of ideas of the "Aha!" variety. He's retiring at the end of the semester, and leaving Buffalo to pursue his research back home in Berkeley. So, I strongly urge you to attend! ======================================================================== CENTER FOR COGNITIVE SCIENCE University at Buffalo, State University of New York Wednesday, April 20, 2005 2:00 pm - 4:00 pm 280 Park Hall, North Campus Leonard Talmy, Ph.D. Department of Linguistics University at Buffalo "THE ATTENTION SYSTEM OF LANGUAGE" This talk reports on work in progress to outline the fundamental attentional system of language. This system includes some fifty basic factors, the "building blocks" of the system. Each factor involves a particular linguistic mechanism that increases or decreases attention on a certain type of linguistic entity. Although able to act alone, the basic factors also regularly combine and interact to produce further attentional effects. This attentional system shows commonalities and differences across individual languages, across modalities (spoken vs. signed language), and across cognitive systems (e.g., between language and visual perception). The methodology used in the analysis, introspection, is itself made the subject of investigation to determine its profile of better and worse function and its consequent relation to other methodologies. A hardcopy of this flyer can be found here: http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/talmyannounce.pdf Please print it out and post it in your department/office. Center for Cognitive Science University at Buffalo, State University of New York 201 Bell Hall, Buffalo, NY 14260 Phone: (716) 645-2177 ext. 795, Fax: (716) 645-3825 Email: ccs-cogsci-info@buffalo.edu Open to the Public. Refreshments will be served. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Apr 12 08:10:33 2005 for ; Tue, 12 Apr 2005 08:10:32 -0400 (EDT) for ; Tue, 12 Apr 2005 08:10:32 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 12 Apr 2005 08:09:35 -0400 12:09:35 -0000 j3CC9Ye2024164; Tue, 12 Apr 2005 08:09:34 -0400 (EDT) -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Tue, 12 Apr 2005 08:09:34 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Center for Cognitive Science Distinguished Speaker: Dedre Gentner Comments: To: lai688@cse.Buffalo.EDU, snerg@cse.Buffalo.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6005 CENTER FOR COGNITIVE SCIENCE University at Buffalo State University of New York DISTINGUISHED SPEAKER SERIES Thursday, April 28, 2005 3:30 pm - 5:00 pm Baird Concert Hall North Campus Dedre Gentner, Ph.D. Department of Psychology Cognitive Science Program Department of Education and Social Policy Northwestern University "Why we're so smart" Human cognitive abilities are remarkable, and even more remarkable is the the rapidity with which children develop cognitive insight. How does this insight arise? A pervasive view in cognitive development is that these rapid gains can only be explained by assuming that infants begin with substantial amounts of innate knowledge. In this talk I propose an alternative approach, centered on mechanisms of human learning. I suggest two powerful forces that contribute to human learning and reasoning ability: (1) analogical processing; and (2) the acquisition of relational language. I will present evidence that the structure-mapping processes that occur during analogy and similarity are a core mechanism by which abstract knowledge arises from experience. Our studies of learning in adults and children show that analogical comparison processes foster learning in several ways: by aligning common relational structure, by suggesting inferences between situations, by focusing attention on relevant differences, and by inviting relational abstractions. A further contributor to human learning and reasoning is the acquisition of relational language. Relational language provides labels that preserve and systematize the relations discovered through comparison processes. It also acts to invite analogical comparisons that reveal common structure. In sum, I suggest that mutual bootstrapping between structure-mapping processes and relational language is a major contributor to human cognition. About Dedre Gentner: Dedre Gentners research is on the psychology of learning and reasoning and the development of cognition and language. Her early work on causal mental models and on the development of word meaning have been influential in cognitive research. Her most important contribution is the structure-mapping theory of analogy and similarity and its implications, including a computational model of similarity processing; a theoretical framework for analogy and metaphor; the evidence for disassociation between the kind of similarity that governs memory retrieval and the kind of similarity that governs on-line mapping and inference. In her developmental work she has proposed a relational shift in childrens similarity processing and has found evidence that this shift is knowledge-driven, rather than maturational. She has also proposed and tested a progressive alignment mechanism whereby comparison processes in ordinary experience can yield theoretical insight. In language learning, Gentners hypothesis of a language-universal advantage for nouns in childrens early word learning that has engendered considerable research. Her recent work unites analogical thinking and language learning and investigates possible interactions between language and cognition. Her theoretical and empirical work provides evidence that relational language has a formative role in the development of relational thought. She is also investigating the hypothesis that analogical processes are integral to language acquisition and use. Open to the Public Free Admission Center for Cognitive Science University at Buffalo 1025 Clemens Hall Buffalo, NY 14260 Phone: (716) 645-2177 ext 795 Fax: (716) 645-3464 http://www.cogsci.buffalo.edu/People/Distinguishedspeakers/gentner05.htm#top From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Apr 12 08:17:02 2005 for ; Tue, 12 Apr 2005 08:17:01 -0400 (EDT) for ; Tue, 12 Apr 2005 08:17:01 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 12 Apr 2005 08:16:23 -0400 for ; Tue, 12 Apr 2005 08:16:22 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Tue, 12 Apr 2005 08:16:22 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Tue, 12 Apr 2005 08:16:22 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: no office hours today To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6007 I won't be in for office hours today. If anyone really needs to talk to me, I'll be available by email, or you can talk to Albert, or set up an appointment with me for later in the week. I do expect to be back tomorrow. ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue Apr 12 12:02:39 2005 for ; Tue, 12 Apr 2005 12:02:38 -0400 (EDT) for ; Tue, 12 Apr 2005 12:02:38 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 12 Apr 2005 12:02:16 -0400 for ; Tue, 12 Apr 2005 12:02:15 -0400 (EDT) -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Tue, 12 Apr 2005 12:02:15 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Code in project report To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: Junk X-UID: 6039 Hello All, A student brought up a good point in my monday recitation: since your code is probably at least 1,000 lines for the automated theorem prover, you should *not* include it as an appendix to your report. I will have an electronic copy of the code and that will do just fine. If you are following Dr. Rapaport's "How to Write" guide, please ignore the section that says to include code. Also ignore the fact that the sample CVA paper of mine that Dr. Rapaport posted on the syllabus contains code as an appendix :) ... I think the spirit of the suggestion is to include code for shorter programs and make the report self-contained...but the code will probably be longer than the body of text for this project. Sorry for any confusion, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 13 09:30:29 2005 for ; Wed, 13 Apr 2005 09:30:29 -0400 (EDT) for ; Wed, 13 Apr 2005 09:30:29 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 13 Apr 2005 09:03:56 -0400 j3DD3tso017834 for ; Wed, 13 Apr 2005 09:03:55 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Wed, 13 Apr 2005 09:03:55 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Wed, 13 Apr 2005 09:03:55 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CSE colloq relevant to KRR To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6118 ------------------------------------------------------------------------ Subject: CSE 4/563: CSE colloq relevant to KRR ------------------------------------------------------------------------ This week's CSE colloquium, while primarily about databases, is highly relevant to KRR, especially because of its overlap with SNePS's belief revision mechanism. ======================================================================== Department of Computer Science and Engineering Presents Leopoldo Bertossi, Carleton University Consistent Answers from Integrated Data Sources When data sources are integrated into a single virtual information system, inconsistencies wrt global integrity constraints are likely to occur because no global maintenance mechanism is available and the souces are autonomous and independent from each other. In consequence, it becomes relevant to retrieve from such a system -at query time- the information that is consistent wrt the global integrity constraints. In this talk, we present a precise definition of consistent answer to a global query in the context of the local-as-view approach to virtual data integration. Furthermore, an algorithm is presented for retrieving from such a mediator-based system the certain answers to monotone queries -extending the available mechanisms for plan generation- and the consistent answers to first order queries. Thursday, April 14, 2005 3:30 - 4:30 PM 330 Student Union Room University at Buffalo - North Campus This talk is free and open to the public Refreshments after the talk in 242 Bell Hall For more information, please email cse-dept@cse.buffalo.edu From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 13 11:50:46 2005 for ; Wed, 13 Apr 2005 11:50:46 -0400 (EDT) for ; Wed, 13 Apr 2005 11:50:45 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 13 Apr 2005 11:17:31 -0400 15:17:31 -0000 j3DFHUbA007970 for ; Wed, 13 Apr 2005 11:17:30 -0400 (EDT) 11:17:30 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Wed, 13 Apr 2005 11:17:30 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: submitting your project To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: Junk X-UID: 6137 I will go over this in recitation...but in case you miss it... To submit your project you should use the "submit" facility on the solaris cse machines. If you wrote your program on a non-cse machine, then copy (i.e., ftp) your files over to a cse machine to submit them. You can submit your files as follows If you are registered for CSE463 then type this at the prompt: submit_cse463 ... If you are registered for CSE563 then type: submit_cse563 ... If you have more than a few files in your project (more than 5) I would suggest that you zip up your files and just submit the zipped archive. Be sure to include: *A README file with instructions on how to compile and run your program (including the environment in which you compiled it). Or a makefile. *Your report...even though I will have a hard copy of it, submit it electronically just in case. *Your well-commented source code files. *(Optional) Any test program that you might have written for your automated theorem prover...or any sample scripts of your running program. Let me know if you have any questions, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 14 13:57:55 2005 for ; Thu, 14 Apr 2005 13:57:55 -0400 (EDT) for ; Thu, 14 Apr 2005 13:57:55 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 14 Apr 2005 13:51:27 -0400 17:51:09 -0000 j3EHp9so023607 for ; Thu, 14 Apr 2005 13:51:09 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 14 Apr 2005 13:51:09 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 14 Apr 2005 13:51:09 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CITATION FOR CHANG AND LEE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6233 A student writes: | Do ... you have a better citation for Chang and Lee? I'm | currently citing it as being from "class notes." The only real | reference I found is a book they wrote together _Symbolic Logic and | Mechanical Theorem Proving_ in 97. | | If it's on the web site somewhere, shame on me, but I didn't see it in | the FOL section... Take a look at: http://www.cse.buffalo.edu/~rapaport/663/F03/unification.pdf :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 14 14:24:53 2005 for ; Thu, 14 Apr 2005 14:24:53 -0400 (EDT) for ; Thu, 14 Apr 2005 14:24:53 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 14 Apr 2005 14:23:59 -0400 j3EINvso023777 for ; Thu, 14 Apr 2005 14:23:57 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 14 Apr 2005 14:23:57 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 14 Apr 2005 14:23:57 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: FORMAL METHODS IN SOFTWARE ENGINEERING To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% ------------------------------------------------------------------------ Subject: CSE 4/563: FORMAL METHODS IN SOFTWARE ENGINEERING ------------------------------------------------------------------------ This may turn out to be a completely irrelevant (to KRR!) talk even if it turns out to be interesting, but since FOL and automated theorem proving are one of the standard techniques (two of the standard techniques?) for software development and verification, I thought I should alert you to this. | Department of Computer Science and Engineering | | Presents | | Formal Aids for the Nurture of Software Systems | | Mathai Joseph, Tata Research Development & Design Centre | A Division of Tata Consultancy Services | | | The use of formal techniques has for a long time been focused on | relatively small and complex applications. The hardware domain lends | itself well to this and it has therefore been the target of some of the | most significant applications of formal techniques. The software | applications that have typically been considered were for small, | safety-critical systems. | | This restricted focus was understandable and necessary while formal | techniques were evolving and practical considerations limited the size | of the system that could be specified and verified. However, there are | now compelling demands for the use of more precise techniques for a | variety of large-scale applications, ranging from smart cards to | financial systems. | | So there are now new reasons to extend the use of formal methods for all | phases of software development: from requirements and software modeling | to coding and testing. Problems of scale still remain so it is important | to focus the use of formal techniques areas where their impact will be | most important. | | Different formal techniques can be used for solving different problems. | For example, use of model-checking during requirements modeling can | identify incomplete or inconsistent specifications, while use of | transformational techniques can be very effective for software modeling | and enable generation of code directly from models. Program analysis | techniques can be used to generate tests that will greatly improve | functional coverage during testing. | | The use of formal techniques continues during software maintenance | through the following kinds of activities: | | a. Remedial: correction of errors discovered during use; | | b. Adaptive: making changes to cater to changes in the operating | environment; | | c. Enhancing: adding new features or capabilities; and | | d. Improving: making the software more robust and easier to maintain. | | It is estimated that the cost of software maintenance amounts to as much | as 90% of the life-cycle cost of a software system. While this calls for | major improvements in maintenance techniques, changes in software | development methods can also help to reduce the need for, and therefore | the cost of, making remedial improvements (i.e. bug fixing). | | In this talk, I will describe the use of formal techniques for different | areas of the software life-cycle and relate this to evidence obtained | through the analysis of a large number of actual software development | and maintenance projects. | | Mathai Joseph joined TCS in 1997 and is Executive Director of Tata | Research Development and Design Centre, the R&D division of Tata | Consultancy Services. He is responsible for setting the research | directions, monitoring progress and converting TCS R&D results into | industrially applicable tools and products. He is responsible for taking | the TRDDC technology tools to market. He has a Ph.D. from the | University of Cambridge, U.K., and worked at the Tata Institute of | Fundamental Research for many years. At various times he has been | Visiting Scientist, INRIA, Rocquencourt (1974), Visiting Professor, at | the Department of Computer Science, Carnegie-Mellon University and | Visiting Professor, Department of Mathematics and Computer Science, | Eindhoven Technical University (1990 - 1992). He is now Visiting | Professor at the University of York, U.K., and Chairman of the Board of | the International Institute for Software Technology. From 1985--97, he | had a Chair in Software Engineering at the Department of Computer | Science, University of Warwick, U.K. where he lead a research group | working on the use of formal methods for safety-critical systems. | | | Friday, April 22, 3005 3:00 - 4:00 PM | 330 SU - University at Buffalo - North Campus | | Refreshments after the talk in 224 Bell Hall For more information email | cse-dept@cse.buffalo.edu From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 14 14:53:34 2005 for ; Thu, 14 Apr 2005 14:53:34 -0400 (EDT) for ; Thu, 14 Apr 2005 14:53:34 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 14 Apr 2005 14:53:03 -0400 j3EIr1so024125 for ; Thu, 14 Apr 2005 14:53:01 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 14 Apr 2005 14:53:01 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 14 Apr 2005 14:53:01 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MOORE'S PROBLEM FOR PROJECT 1 (again?) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% This did not seem to get posted the first time I sent it. | ------------------------------------------------------------------------ | Subject: CSE 4/563: MOORE'S PROBLEM FOR PROJECT 1 | ------------------------------------------------------------------------ | | A student writes: | | | Where can I find it? I couldn't find a reference on the project listing or | | your web site. | | The link is on the html version of the project, but here's a direct | link: | | http://www.cse.buffalo.edu/~rapaport/563S05/folaskrr.html#moore | | (well, actually, it's the direct link from the html version of the | project to yet another webpage that has a direct link to the pdf version | of the article :-) | | PS: The html version of the project is at: | http://www.cse.buffalo.edu/~rapaport/563S05/Proj1/proj1/proj1.html | From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 14 15:11:49 2005 for ; Thu, 14 Apr 2005 15:11:48 -0400 (EDT) for ; Thu, 14 Apr 2005 15:11:48 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 14 Apr 2005 14:56:24 -0400 18:56:24 -0000 j3EIuNso024165 for ; Thu, 14 Apr 2005 14:56:23 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 14 Apr 2005 14:56:23 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Thu, 14 Apr 2005 14:56:23 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: MOORE'S PROBLEM FOR PROJECT 1 To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% X-Keywords: X-UID: 6237 | > The link is on the html version of the project, but here's a direct | > link: | > | > http://www.cse.buffalo.edu/~rapaport/563S05/folaskrr.html#moore | | It's password protected. I no longer have that password. For those of you who have forgotten :-) login=CSE563 pw=Spring2005 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Apr 15 09:38:36 2005 for ; Fri, 15 Apr 2005 09:38:36 -0400 (EDT) for ; Fri, 15 Apr 2005 09:38:36 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 15 Apr 2005 09:32:04 -0400 13:32:04 -0000 j3FDW3so026657 for ; Fri, 15 Apr 2005 09:32:03 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Fri, 15 Apr 2005 09:32:03 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Fri, 15 Apr 2005 09:32:03 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 9 ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% HW 9 ANSWERS ------------------------------------------------------------------------ 1. a) {y/c, z/b, x/c} 0 = missing 1 = incorrect MGU 2 = incorrect MGU, work shown 3 = correct MGU ------------------------------------------------------------------------ b) fails: occurs check fails when trying to unify b and c (in final step, the DS = {b,c}, but in that DS there's no variable in it that occurs in a term in it) 0123 points ------------------------------------------------------------------------ c) This one is worth working out in detail. Following our unification algorithm and using some obvious abbreviations, we have: k=0 W0 = {Y(s(x),x), Y(s(y),b)} theta0 = {} W0 not singleton DS0 = {x,y} At this point, we have two choices for theta1; I'll consider them in parallel columns: theta1 = {}o{x/y} = {x/y} theta1 = {}o{y/x} = {y/x} W1={Y(s(y),y), Y(s(y),b)} W1={Y(s(x),x), Y(s(x),b)} Note that both W1s are essentially the same wff, modulo change of bound variable! k=1 W1 not singleton DS1 = {y,b} DS1 = {x,b} theta2 = {x/y}o{y/b} theta2 = {y/x}o{x/b} = {x/b, y/b} = {y/b, x/b} You need to apply the full definition of composition of substitutions to see why theta2 is what it is. Note that no matter which substitution was chosen for theta1, theta2 is identical in both cases! W2 = {Y(s(b),b), Y(s(b),b)} ditto! W2 is a singleton, so we're done. theta2 is MGU. 0123 points ------------------------------------------------------------------------ d) fails; x occurs in son-of(x) 0123 points Total for #1 = 12 points ------------------------------------------------------------------------ 2. Answers may vary, depending on how the conclusion is translated. (E.g., another translation is: AyAz[(Horse(z) ^ Head(y,z)) > Ew[Animal(w) ^ Head(y,w)]] ) a) Ax[Horse(x) > Animal(x)] |- Ay[Ez[Horse(z) ^ Head(y,z)] > Ew[Animal(w) ^ Head(y,w)]] (see: http://www.cse.buffalo.edu/~rapaport/Papers/logic,predicate.pdf p. 870) premise: 0123 conclusion: 0123 total = 6 points ------------------------------------------------------------------------ b) CF(premise) = [-Horse(x), Animal(x)] CF(-conclusion): -Ay[Ez[Horse(z) ^ Head(y,z)] > Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[-Ez[Horse(z) ^ Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[Az.-[Horse(z) ^ Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[Az[-Horse(z) v -Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> Ey.-[Az[-Horse(z) v -Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> Ey[-Az[-Horse(z) v -Head(y,z)] ^ -Ew[Animal(w) ^ Head(y,w)]] ---> Ey[Az.-[-Horse(z) v -Head(y,z)] ^ Aw.-[Animal(w) ^ Head(y,w)]] ---> Ey[Az[Horse(z) ^ Head(y,z)] ^ Aw[-Animal(w) v -Head(y,w)]] ---> EyAzAw[(Horse(z) ^ Head(y,z)) ^ (-Animal(w) v -Head(y,w))] ---> AzAw[(Horse(z) ^ Head(c,z)) ^ (-Animal(w) v -Head(c,w))], where c is a Skolem constant ---> (Horse(z) ^ Head(c,z)) ^ (-Animal(w) v -Head(c,w)) ---> {[Horse(z)], [Head(c,z)], [-Animal(w), -Head(c,w)]} premise: 0123 conclusion: 0123 total = 6 points ------------------------------------------------------------------------ c) {[-Horse(x), Animal(x)], (1) [Horse(z)], (2) [Head(c,z)], (3) [-Animal(w), -Head(c,w)]} (4) resolve 1,2 with MGU = {x/z} to get [Animal(z)] (5) resolve 5,4 with MGU = {w/z} to get [-Head(c,z)] (6) resolve 6,3 to get [] proof: 0123 MGUs: 0123 total = 6 points subtotal for #2 = 18 points ------------------------------------------------------------------------ 3. a) i) Ax[P(x) > N(x)] ii) P(d) iii) -Ex[N(x) ^ W(x)] or: Ax[N(x) > -W(x)] iv) -W(d) 01234 (i.e., 1 point per sentence) ------------------------------------------------------------------------ b) i) [-P(x), N(x)] ii) [P(d)] iii) [-N(x), -W(x)] iv) [W(d)] 01234 (i.e., 1 point per sentence) ------------------------------------------------------------------------ c) resolve iii, iv with MGU = {x/d} to get [-N(d)] (v) resolve i, v with MGU = {x/d} to get [-P(d)] (vi) resolve vi, ii to get [] proof: 0123 MGUs: 0123 subtotal = 6 points total for #3 = 14 points ------------------------------------------------------------------------ 4. a) A) For every natural number, there is a(nother) natural number such that: the first is greater than or equal to the second. 0123 points B) There is a (particular) natural number such that: every natural number is greater than or equal to it. 0123 points subtotal = 6 points ------------------------------------------------------------------------ b) Yes, A is true under this interpretation... 01 point ... because you can take the two numbers to be the same! 0123 points subtotal = 4 points ------------------------------------------------------------------------ c) Yes, B, A is true under this interpretation... 01 point ... because you can let 0 be the "particular natural number" 0123 points subtotal = 4 points ------------------------------------------------------------------------ Below, I will use "GEQ" as a prefix predicate instead of the infix predicate ">=", which, first, is hard to type in ASCII and, second, might be confused with the material implication symbol. d) Try to show that A |- B: CF(A): AxEy.GEQ(x,y) ---> Ax.GEQ(x,f(x)) ---> [GEQ(x,f(x))] (1) CF(-B): -EyAx.GEQ(x,y) ---> -EyAz.GEQ(z,y) (rewriting bound variables) ---> AyEz.-GEQ(z,y) ---> Ay.-GEQ(g(y),y) ---> [-GEQ(g(y),y)] (2) where f(x), g(y) are distinct Skolem functions. Compute MGU of GEQ(x,f(x)), GEQ(g(y),y): theta0 = {} DS0 = {x, g(y)} occurs check is true: x is a variable that does not occur in the term g(y) theta1 = {}o{x/g(y)} = {x/g(y)} W1 = {GEQ(g(y), f(g(y))), GEQ(g(y), y)} DS1 = {y, f(g(y))} occurs check fails: y is the only variable in the DS, but it occurs in the term f(g(y)) Therefore, the two clauses cannot be resolved! Therefore, can't show that A |- B 0123 points ------------------------------------------------------------------------ e) Try to show that B |- A: CF(B): EyAx.GEQ(x,y) ---> Ax.GEQ(x,c) ---> [GEQ(x,c)] (1) CF(-A): -AxEy.GEQ(x,y) ---> -AzEy.GEQ(z,y) ---> EzAy.-GEQ(z,y) ---> Ay.-GEQ(d,y) ---> [-GEQ(d,y)] (2) where c,d are distinct Skolem constants. resolve 1,2 with MGU = {x/d, y/c} to get [] 0123 points total for #4 = 20 points ------------------------------------------------------------------------ Grand Total = 64 points Letter 463 both 563 ------ --- ---- --- A 61-64 A- 58-60 B+ 54-57 B 51-53 B- 47-50 C+ 44-46 C 37-43 22-43 C- 29-36 D+ 22-28 D 12-21 F 0-11 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 17 14:05:14 2005 for ; Sun, 17 Apr 2005 14:05:13 -0400 (EDT) for ; Sun, 17 Apr 2005 14:05:13 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 17 Apr 2005 13:56:30 -0400 17:56:29 -0000 for ; Sun, 17 Apr 2005 13:56:25 -0400 (EDT) -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Sun, 17 Apr 2005 13:56:25 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: CSE 4/563: HW 9 ANSWERS (fwd) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% One small note in about the answers posted for hw #9. The answer for #4 parts (d) and (e) were swapped. These were the two exercises where you had to show A |- B and B |- A. Part (d) said to "try to prove that A follows from B" which is B |- A. Part (e) said to "try to prove that B follows from A" which is A |- B. The language "conclusion follows from premise" has the reverse syntax of the notation for |-. Anyway...the answers are correct...just swap the part labels as below ------------------------------------------------------------------------ (was part d) e) Try to show that A |- B: CF(A): AxEy.GEQ(x,y) ---> Ax.GEQ(x,f(x)) ---> [GEQ(x,f(x))] (1) CF(-B): -EyAx.GEQ(x,y) ---> -EyAz.GEQ(z,y) (rewriting bound variables) ---> AyEz.-GEQ(z,y) ---> Ay.-GEQ(g(y),y) ---> [-GEQ(g(y),y)] (2) where f(x), g(y) are distinct Skolem functions. Compute MGU of GEQ(x,f(x)), GEQ(g(y),y): theta0 = {} DS0 = {x, g(y)} occurs check is true: x is a variable that does not occur in the term g(y) theta1 = {}o{x/g(y)} = {x/g(y)} W1 = {GEQ(g(y), f(g(y))), GEQ(g(y), y)} DS1 = {y, f(g(y))} occurs check fails: y is the only variable in the DS, but it occurs in the term f(g(y)) Therefore, the two clauses cannot be resolved! Therefore, can't show that A |- B 0123 points ------------------------------------------------------------------------ (was part e) d) Try to show that B |- A: CF(B): EyAx.GEQ(x,y) ---> Ax.GEQ(x,c) ---> [GEQ(x,c)] (1) CF(-A): -AxEy.GEQ(x,y) ---> -AzEy.GEQ(z,y) ---> EzAy.-GEQ(z,y) ---> Ay.-GEQ(d,y) ---> [-GEQ(d,y)] (2) where c,d are distinct Skolem constants. resolve 1,2 with MGU = {x/d, y/c} to get [] 0123 points total for #4 = 20 points ------------------------------------------------------------------------ Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 17 14:38:05 2005 for ; Sun, 17 Apr 2005 14:38:05 -0400 (EDT) for ; Sun, 17 Apr 2005 14:38:05 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 17 Apr 2005 14:35:10 -0400 j3HIZ9e2023612; Sun, 17 Apr 2005 14:35:09 -0400 (EDT) -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Sun, 17 Apr 2005 14:35:09 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: CSE 4/563: HW 9 ANSWERS (fwd) Comments: To: ag33@cse.Buffalo.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% | Date: Sun, 17 Apr 2005 13:56:25 -0400 | From: Albert Goldfain | Subject: CSE 4/563: HW 9 ANSWERS (fwd) | | One small note in about the answers posted for hw #9. The answer for #4 | parts (d) and (e) were swapped. oops :-| From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun Apr 17 14:51:01 2005 for ; Sun, 17 Apr 2005 14:51:01 -0400 (EDT) for ; Sun, 17 Apr 2005 14:51:01 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 17 Apr 2005 14:50:10 -0400 18:50:10 -0000 for ; Sun, 17 Apr 2005 14:50:00 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Sun, 17 Apr 2005 14:50:00 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Sun, 17 Apr 2005 14:50:00 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: UPDATED SYLLABUS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% ------------------------------------------------------------------------ Subject: CSE 4/563: UPDATED SYLLABUS ------------------------------------------------------------------------ I've updated the syllabus again, to reflect some recent changes and to add some new recommended readings. From rapaport@cse.Buffalo.EDU Mon Apr 18 16:49:16 2005 for ; Mon, 18 Apr 2005 16:49:16 -0400 (EDT) for ; Mon, 18 Apr 2005 16:49:16 -0400 (EDT) Mon, 18 Apr 2005 16:49:15 -0400 (EDT) Mon, 18 Apr 2005 16:49:15 -0400 (EDT) Date: Mon, 18 Apr 2005 16:49:15 -0400 (EDT) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu, ddechert@buffalo.edu, mclee2@buffalo.edu, mwkibby@buffalo.edu, nbailey@buffalo.edu, rapaport@cse.Buffalo.EDU, shuh@buffalo.edu, solanax73@yahoo.com, tmadigan@buffalo.edu, tmchrist@buffalo.edu Subject: Dedre Gentner: CogSci Distinguished Speaker CENTER FOR COGNITIVE SCIENCE University at Buffalo State University of New York DISTINGUISHED SPEAKER SERIES Thursday, April 28, 2005 3:30 pm - 5:00 pm Baird Concert Hall North Campus Dedre Gentner, Ph.D. Department of Psychology Cognitive Science Program Department of Education and Social Policy Northwestern University "Why we're so smart" Human cognitive abilities are remarkable, and even more remarkable is the the rapidity with which children develop cognitive insight. How does this insight arise? A pervasive view in cognitive development is that these rapid gains can only be explained by assuming that infants begin with substantial amounts of innate knowledge. In this talk I propose an alternative approach, centered on mechanisms of human learning. I suggest two powerful forces that contribute to human learning and reasoning ability: (1) analogical processing; and (2) the acquisition of relational language. I will present evidence that the structure-mapping processes that occur during analogy and similarity are a core mechanism by which abstract knowledge arises from experience. Our studies of learning in adults and children show that analogical comparison processes foster learning in several ways: by aligning common relational structure, by suggesting inferences between situations, by focusing attention on relevant differences, and by inviting relational abstractions. A further contributor to human learning and reasoning is the acquisition of relational language. Relational language provides labels that preserve and systematize the relations discovered through comparison processes. It also acts to invite analogical comparisons that reveal common structure. In sum, I suggest that mutual bootstrapping between structure-mapping processes and relational language is a major contributor to human cognition. About Dedre Gentner: Dedre Gentners research is on the psychology of learning and reasoning and the development of cognition and language. Her early work on causal mental models and on the development of word meaning have been influential in cognitive research. Her most important contribution is the structure-mapping theory of analogy and similarity and its implications, including a computational model of similarity processing; a theoretical framework for analogy and metaphor; the evidence for disassociation between the kind of similarity that governs memory retrieval and the kind of similarity that governs on-line mapping and inference. In her developmental work she has proposed a relational shift in childrens similarity processing and has found evidence that this shift is knowledge-driven, rather than maturational. She has also proposed and tested a progressive alignment mechanism whereby comparison processes in ordinary experience can yield theoretical insight. In language learning, Gentners hypothesis of a language-universal advantage for nouns in childrens early word learning that has engendered considerable research. Her recent work unites analogical thinking and language learning and investigates possible interactions between language and cognition. Her theoretical and empirical work provides evidence that relational language has a formative role in the development of relational thought. She is also investigating the hypothesis that analogical processes are integral to language acquisition and use. Open to the Public Free Admission Center for Cognitive Science University at Buffalo 1025 Clemens Hall Buffalo, NY 14260 Phone: (716) 645-2177 ext 795 Fax: (716) 645-3464 http://www.cogsci.buffalo.edu/People/Distinguishedspeakers/gentner05.htm#top From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 18 20:40:00 2005 for ; Mon, 18 Apr 2005 20:40:00 -0400 (EDT) for ; Mon, 18 Apr 2005 20:39:59 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 18 Apr 2005 20:39:38 -0400 00:39:38 -0000 j3J0dbbA006630 for ; Mon, 18 Apr 2005 20:39:37 -0400 (EDT) 20:39:37 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Mon, 18 Apr 2005 20:39:37 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Sneps representation of "Mother/Daughter reading together" example To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% Hello all, If you are in the Wed. representation and want to be surprised by this example then don't look :) The following is my interaction with SNePS in representing "The mother and daughter read the book together"...I am skipping all of the nice object-propername and lex arcs to get to the heart of the uniqueness issue. My comments are after the semicolons like this: ;;;comment ---------------------------------------------------------------------- Type `(sneps)' or `(snepslog)' to get started. cl-user(2): (sneps) Welcome to SNePS-2.6 [PL:1a 2004/08/26 23:05:27] Copyright (C) 1984--2004 by Research Foundation of State University of New York. SNePS comes with ABSOLUTELY NO WARRANTY! Type `(copyright)' for detailed copyright information. Type `(demo)' for a list of example applications. 4/18/2005 20:15:13 ;;; ;;;Define the necessary relations. Act and action are already defined, ;;;but it won't hurt to redefine them. ;;; * (define agent act action object) act is already defined. action is already defined. (agent act action object) CPU time : 0.00 ;;; ;;; Assert that the mother read the book ;;; * (describe (assert agent mother act (build action read object book))) (m2! (act (m1 (action read) (object book))) (agent mother)) (m2!) CPU time : 0.00 ;;; ;;; Assert that the daughter read the book (taking part in exactly the ;;; same act as the mother) ;;; * (describe (assert agent daughter act (build action read object book))) (m3! (act (m1 (action read) (object book))) (agent daughter)) (m3!) CPU time : 0.00 ;;; ;;; Notice that m1 is pointed to by *both* act arcs, the "mother's" ;;; and the "daughter's". So even though we told CASSIE ;;; (build action read book) twice above (build is the SNePSUL way of ;;; introducing unasserted molecular nodes in SNePS), since it refers to ;;; the same piece of network she will only build it once! (this is one ;;; aspect of the uniqueness principle in SNePS). ;;; * (describe *nodes) (m3! (act (m1 (action read) (object book))) (agent daughter)) (m2! (act (m1)) (agent mother)) (m3! daughter m2! m1 book read mother) CPU time : 0.00 * (lisp) "End of SNePS" cl-user(3): :exit ; Exiting pollux {~} > exit exit script done on Mon 18 Apr 2005 08:17:17 PM EDT Therefore...to finish answering the question raised in recitation: "Is there a difference between my representation of 'The mother and daughter read the book together' and the sentence 'The mother read the book and the daughter read the book'?" The answer is NO! We don't need to add an andor connective here b/c (as I said) all asserted propositions are implicitly conjoined in SNePS. We can never build the following network in SNePS: ;;; ;;;Impossible...violates uniqueness principle ;;; (m3! (act (m1 (action read) (object book))) (agent daughter)) (m2! (act (m4 (action read) (object book))) (agent mother)) because m1 and m4 are identical (and thus must be a single node by the uniqueness principle). Fantastic question though :) Albert ps...make sure you do those SNePS readings! From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 18 20:54:09 2005 for ; Mon, 18 Apr 2005 20:54:09 -0400 (EDT) for ; Mon, 18 Apr 2005 20:54:09 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 18 Apr 2005 20:50:23 -0400 00:50:23 -0000 j3J0oMbA008175 for ; Mon, 18 Apr 2005 20:50:22 -0400 (EDT) 20:50:22 -0400 (EDT) X-PM-Spam-Prob: : 7% Date: Mon, 18 Apr 2005 20:50:22 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: demo'ing your work To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-PM-Spam-Prob: : 7% Instruction #2 of Programming Project 2 says: "2. Do Sections 1-7 of the tutorial (Section 1 is instructions on running SNePS). Save all your interactions in one or more files." One way to save these interactions (for steps 2 and 3) is to create a "demo" file. Essentially you put all of the commands you would normally type at the SNePSUL prompt (*) in a file named ".demo". Then to "run" this file from within SNePSUL just type the following at the prompt: (demo ".demo") this will have the effect of going through all of your commands one by one. If you want to slow this down so that you can see what happens with each command, you can force a pause after each command by typing the following instead: (demo ".demo" :av) These instructions are actually in the tutorial...section 1.2 http://www.cse.buffalo.edu/sneps/Tutorial/node3.html Albert From rapaport@cse.Buffalo.EDU Wed Apr 20 08:00:45 2005 Wed, 20 Apr 2005 08:00:45 -0400 (EDT) for ; Wed, 20 Apr 2005 08:00:43 -0400 (EDT) Wed, 20 Apr 2005 08:00:42 -0400 (EDT) Wed, 20 Apr 2005 08:00:42 -0400 (EDT) Date: Wed, 20 Apr 2005 08:00:42 -0400 (EDT) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu, snerg@cse.Buffalo.EDU Subject: Talmy's Last Talk (at UB, before retirement :-) TODAY! TODAY! TODAY! TODAY! TODAY! TODAY! TODAY! TODAY! TODAY! CENTER FOR COGNITIVE SCIENCE University at Buffalo, State University of New York Wednesday, April 20, 2005 2:00 pm - 4:00 pm 280 Park Hall, North Campus Leonard Talmy, Ph.D. Department of Linguistics University at Buffalo "THE ATTENTION SYSTEM OF LANGUAGE" This talk reports on work in progress to outline the fundamental attentional system of language. This system includes some fifty basic factors, the "building blocks" of the system. Each factor involves a particular linguistic mechanism that increases or decreases attention on a certain type of linguistic entity. Although able to act alone, the basic factors also regularly combine and interact to produce further attentional effects. This attentional system shows commonalities and differences across individual languages, across modalities (spoken vs. signed language), and across cognitive systems (e.g., between language and visual perception). The methodology used in the analysis, introspection, is itself made the subject of investigation to determine its profile of better and worse function and its consequent relation to other methodologies. A hardcopy of this flyer can be found here: http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/talmyannounce.pdf Please print it out and post it in your department/office. Center for Cognitive Science University at Buffalo, State University of New York 201 Bell Hall, Buffalo, NY 14260 Phone: (716) 645-2177 ext. 795, Fax: (716) 645-3825 Email: ccs-cogsci-contact@buffalo.edu Open to the Public. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed Apr 20 08:28:35 2005 for ; Wed, 20 Apr 2005 08:28:35 -0400 (EDT) for ; Wed, 20 Apr 2005 08:28:35 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 20 Apr 2005 08:18:37 -0400 12:18:37 -0000 for ; Wed, 20 Apr 2005 08:18:36 -0400 (EDT) -0400 (EDT) Date: Wed, 20 Apr 2005 08:18:36 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Re: Talmy's Last Talk (at UB, before retirement :-) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU In-Reply-To: <200504201200.j3KC0gpU023371@castor.cse.Buffalo.EDU> I "second" Dr. Rapaport's urging you to see Dr. Talmy's talk today. I am taking my third class with Dr. Talmy this semeseter. He is almost always good for an epiphany or two with respect to the workings of the human mind. The "attentional system of language" may not me the most central of issues to KRR, but I think you will all like this talk. Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 21 20:43:54 2005 for ; Thu, 21 Apr 2005 20:43:54 -0400 (EDT) for ; Thu, 21 Apr 2005 20:43:53 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 21 Apr 2005 20:43:27 -0400 00:43:27 -0000 for ; Thu, 21 Apr 2005 20:43:27 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 21 Apr 2005 20:43:27 -0400 (EDT) Date: Thu, 21 Apr 2005 20:43:27 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: SYLLABUS UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: SYLLABUS UPDATE ------------------------------------------------------------------------ I have updated the syllabus, hopefully for the last time. There are several new readings. I'll discuss them on Friday. From rapaport@cse.Buffalo.EDU Fri Apr 22 20:54:53 2005 Fri, 22 Apr 2005 20:54:53 -0400 (EDT) Fri, 22 Apr 2005 20:54:53 -0400 (EDT) Date: Fri, 22 Apr 2005 20:54:53 -0400 (EDT) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu, ddechert@buffalo.edu, mclee2@buffalo.edu, mwkibby@buffalo.edu, nbailey@buffalo.edu, rapaport@cse.Buffalo.EDU, shuh@buffalo.edu, solanax73@yahoo.com, tmadigan@buffalo.edu, tmchrist@buffalo.edu Subject: Dedre Gentner Dedre Gentner is giving two talks next week. The first is listed below. I'll send info on the second in my next email. She is this year's Cognitive Science Distinguished Speaker. She is a psychologist who has done a lot of work on various kinds of knowledge representation & reasoning issues (especially analogical reasoning) and on language learning. CENTER FOR COGNITIVE SCIENCE University at Buffalo State University of New York Wednesday, April 27, 2005 2:00 pm - 4:00 pm 280 Park Hall North Campus Dedre Gentner, Ph.D. Department of Psychology Cognitive Science Program Department of Education and Social Policy Northwestern University "Acquiring and Using Relational Representations: Computational and Empirical Details and Theoretical Speculations" Open to the Public Free Admission Center for Cognitive Science University at Buffalo 1025 Clemens Hall Buffalo, NY 14260 Phone: (716) 645-2177 ext 795 Fax: (716) 645-3464 http://www.cogsci.buffalo.edu/Activities/Colloquium/CLLQs05/2005spring.htm From rapaport@cse.Buffalo.EDU Fri Apr 22 20:56:23 2005 Fri, 22 Apr 2005 20:56:23 -0400 (EDT) Fri, 22 Apr 2005 20:56:23 -0400 (EDT) Date: Fri, 22 Apr 2005 20:56:23 -0400 (EDT) From: "William J. Rapaport" To: cse563-sp05-list@listserv.buffalo.edu, ddechert@buffalo.edu, mclee2@buffalo.edu, mwkibby@buffalo.edu, nbailey@buffalo.edu, rapaport@cse.Buffalo.EDU, shuh@buffalo.edu, solanax73@yahoo.com, tmadigan@buffalo.edu, tmchrist@buffalo.edu Subject: Dedre Gentner, talk #2 Here is talk #2, aimed at the general public. A written version of this talk is on the web in PDF at: http://www.psych.nwu.edu/psych/people/faculty/gentner/newpdfpapers/GentnerWW03.pdf ======================================================================== CENTER FOR COGNITIVE SCIENCE University at Buffalo State University of New York DISTINGUISHED SPEAKER SERIES 2005 Thursday, April 28, 2005 3:30 pm - 5:00 pm Baird Concert Hall North Campus Dedre Gentner, Ph.D. Department of Psychology Cognitive Science Program Department of Education and Social Policy Northwestern University "Why we're so smart" Human cognitive abilities are remarkable, and even more remarkable is the the rapidity with which children develop cognitive insight. How does this insight arise? A pervasive view in cognitive development is that these rapid gains can only be explained by assuming that infants begin with substantial amounts of innate knowledge. In this talk I propose an alternative approach, centered on mechanisms of human learning. I suggest two powerful forces that contribute to human learning and reasoning ability: (1) analogical processing; and (2) the acquisition of relational language. I will present evidence that the structure-mapping processes that occur during analogy and similarity are a core mechanism by which abstract knowledge arises from experience. Our studies of learning in adults and children show that analogical comparison processes foster learning in several ways: by aligning common relational structure, by suggesting inferences between situations, by focusing attention on relevant differences, and by inviting relational abstractions. A further contributor to human learning and reasoning is the acquisition of relational language. Relational language provides labels that preserve and systematize the relations discovered through comparison processes. It also acts to invite analogical comparisons that reveal common structure. In sum, I suggest that mutual bootstrapping between structure-mapping processes and relational language is a major contributor to human cognition. About Dedre Gentner: Dedre Gentners research is on the psychology of learning and reasoning and the development of cognition and language. Her early work on causal mental models and on the development of word meaning have been influential in cognitive research. Her most important contribution is the structure-mapping theory of analogy and similarity and its implications, including a computational model of similarity processing; a theoretical framework for analogy and metaphor; the evidence for disassociation between the kind of similarity that governs memory retrieval and the kind of similarity that governs on-line mapping and inference. In her developmental work she has proposed a relational shift in childrens similarity processing and has found evidence that this shift is knowledge-driven, rather than maturational. She has also proposed and tested a progressive alignment mechanism whereby comparison processes in ordinary experience can yield theoretical insight. In language learning, Gentners hypothesis of a language-universal advantage for nouns in childrens early word learning that has engendered considerable research. Her recent work unites analogical thinking and language learning and investigates possible interactions between language and cognition. Her theoretical and empirical work provides evidence that relational language has a formative role in the development of relational thought. She is also investigating the hypothesis that analogical processes are integral to language acquisition and use. Open to the Public Free Admission Center for Cognitive Science University at Buffalo 1025 Clemens Hall Buffalo, NY 14260 Phone: (716) 645-2177 ext 795 Fax: (716) 645-3464 http://www.cogsci.buffalo.edu/People/Distinguishedspeakers/gentner05.htm#top From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon Apr 25 13:06:51 2005 for ; Mon, 25 Apr 2005 13:06:50 -0400 (EDT) for ; Mon, 25 Apr 2005 13:06:50 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 25 Apr 2005 12:44:32 -0400 16:44:29 -0000 j3PGiTso024200 for ; Mon, 25 Apr 2005 12:44:29 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Mon, 25 Apr 2005 12:44:29 -0400 (EDT) Date: Mon, 25 Apr 2005 12:44:29 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 10 DUE DATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU X-Keywords: X-UID: 6598 ------------------------------------------------------------------------ Subject: CSE 4/563: HW 10 DUE DATE ------------------------------------------------------------------------ As announced in class on Friday, HW 10 is now due on Monday, May 2. I've changed the websites to reflect this. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu Apr 28 11:19:16 2005 for ; Thu, 28 Apr 2005 11:19:15 -0400 (EDT) for ; Thu, 28 Apr 2005 11:19:15 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 28 Apr 2005 11:18:36 -0400 15:18:36 -0000 j3SFIajd029434 for ; Thu, 28 Apr 2005 11:18:36 -0400 (EDT) 11:18:36 -0400 (EDT) Date: Thu, 28 Apr 2005 11:18:36 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: submitting project 2 (fwd) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU A student writes: > >Are we supposed to submit project 2 (demo file & annotated .doc file) >using the submit script, or are we just to hand in a hardcopy? > Please submit an online version of your demo file (using submit_cse563 or submit_cse463) *AND* hand in a hard copy of your annotated demo/script in class. Same submission procedure as project 1. Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Apr 29 15:44:27 2005 for ; Fri, 29 Apr 2005 15:44:26 -0400 (EDT) for ; Fri, 29 Apr 2005 15:44:26 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 29 Apr 2005 15:43:55 -0400 19:43:54 -0000 j3TJhsso023399 for ; Fri, 29 Apr 2005 15:43:54 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Fri, 29 Apr 2005 15:43:54 -0400 (EDT) Date: Fri, 29 Apr 2005 15:43:54 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: DIRECTORY UPDATE To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: DIRECTORY UPDATE ------------------------------------------------------------------------ I've updated the directory of documents to add: * Info on the final exam * Course summary (we'll go over it on Monday) * Lecture notes on the "fundamental problem of KRR" From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri Apr 29 21:18:14 2005 for ; Fri, 29 Apr 2005 21:18:14 -0400 (EDT) for ; Fri, 29 Apr 2005 21:18:14 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 29 Apr 2005 21:18:07 -0400 01:18:06 -0000 for ; Fri, 29 Apr 2005 21:18:06 -0400 (EDT) -0400 (EDT) Date: Fri, 29 Apr 2005 21:18:06 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: finishing up To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU first a little humor (at my own pain)... I pulled my back a bit today b/c my bag was so heavy with your papers (proj1 + proj2)...so you can all say that you wrote the paper that broke the TA's back :) I am almost done grading your proj 1's. They will be available to the monday recitation people during recitation (giving them some incentive to come :) and to the wed recitation people right after class on monday. I will not be holding my office hours next week...but if you would like to meet with me then send me an email and we will set up a time. Also...this weekend I will post my method for grading the "Description (part a)" sections of CF, Unification, and Resolution in project 1...the point distribution was the same as the rubric Dr. Rapaport sent around...I will detail what kind of answers got 0,5,10,15. Have a nice weekend, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun May 1 10:13:28 2005 for ; Sun, 1 May 2005 10:13:28 -0400 (EDT) for ; Sun, 1 May 2005 10:13:27 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 1 May 2005 10:13:12 -0400 14:13:12 -0000 for ; Sun, 1 May 2005 10:13:11 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Sun, 1 May 2005 10:13:11 -0400 (EDT) Date: Sun, 1 May 2005 10:13:11 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: PROJECT 1 ANSWERS AND GRADING To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: PROJECT 1 ANSWERS AND GRADING ------------------------------------------------------------------------ 1b) Using obvious abbreviations, and taking things step by step... forall x[Ax > ( Px = exists y[Ay ^ Exy] ) ] ---> forall x[-Ax v ( ) ] ---> forall x[-Ax v ( {Px > exists y[Ay ^ Exy]} ^ {exists y [ Ay ^ Exy ] > Px} ) ] ---> forall x[-Ax v ( {-Px v exists y[Ay ^ Exy]} ^ {-exists y[ Ay ^ Exy ] v Px} ) ] ---> forall x[-Ax v ( {-Px v exists y[Ay ^ Exy]} ^ {forall y.-[Ay ^ Exy] v Px} ) ] ---> forall x[-Ax v ( {-Px v exists y[Ay ^ Exy]} ^ {forall y[ -Ay v -Exy] v Px} ) ] ---> forall x[-Ax v ( {-Px v exists y[Ay ^ Exy]} ^ {forall z[ -Az v -Exz] v Px} ) ] ---> (forall x)(exists y)(forall z)[-Ax v ({-Px v [Ay ^ Exy]} ^ {-Az v -Exz v Px})] ---> (forall x)(forall z)[-Ax v ({-Px v [Af(x) ^ Exf(x)]} ^ {-Az v -Exz v Px})] ---> [-Ax v ({-Px v [Af(x) ^ Exf(x)]} ^ {-Az v -Exz v Px})] ---> this is of the form alpha v (beta ^ gamma), where alpha = -Ax beta = {-Px v [Af(x) ^ Exf(x)]} gamma = {-Az v -Exz v Px} multiplying out, we get (alpha v beta) ^ (alpha v gamma), i.e.: (-Ax v {-Px v [Af(x) ^ Exf(x)]}) ^ (-Ax v {-Az v -Exz v Px}) regrouping, using the associativity of v, we get: ([-Ax v -Px] v [Af(x) ^ Exf(x)]) ^ (-Ax v -Az v -Exz v Px) from here on, the right conjunct won't change (much)... the left conjunct is now of the form alpha v (beta ^ gamma), where alpha = [-Ax v -Px] beta = Af(x) gamma = Exf(x) multiplying out, we get (alpha v beta) ^ (alpha v gamma), all conjoined with the right conjunct, i.e.: [-Ax v -Px v Af(x)] ^ [-Ax v -Px v Exf(x)] ^ [-Ax v -Az v -Exz v Px] Changing variables in each conjunct (a subtle point, often missed, not worth taking off too many points for :-) ... [-Ax v -Px v Af(x)] ^ [-Ax1 v -Px1 v Ex1f(x1)] ^ [-Ax2 v -Az v -Ex2z v Px2] This is now in CNF; rewriting, we get CF: { [-Ax, -Px, Af(x)], [-Ax1, -Px1, Ex1f(x1)], [-Ax2, -Az, -Ex2z, Px2] } ------------------------------------------------------------------------ 2b) (i) W={P(a,x,c), P(y,b,z)} {y/a} W={P(a,x,c), P(a,b,z)} {y/a, x/b} W={P(a,b,c), P(a,b,z)} {y/a, x/b, z/c} W={P(a,b,c)} (ii) W={P(a,x,c), P(y,b,y)} {y/a} W={P(a,x,c), P(a,b,a)} {y/a, x/b} W={P(a,b,c), P(a,b,a)} in the next step, occurs check fails: a & c are both constants, so there's no variable; therefore, NOT UNIFIABLE (iii) W={P(x,x,c), P(u,v,u)} {x/u} (*) W={P(u,u,c), P(u,v,u)} {x/v, u/v} (+) W={P(v,v,c), P(v,v,v)} {x/c, u/c, v/c} (#) W={P(c,c,c)} (*) alteratively, use {u/x} yielding W={P(x,x,c), P(x,v,x)} and continue as above; should give same final result in this particular case. (+) note that composition changes the first member of this set! Note, too, that could have used v/u, yielding multiple possible pathways through the unification algorithm; all should give same final result in this particular case. (#) note again that composition changes the previous members! (iv) W={P(x, f(x) , f(y)), P(f(a), f(z) , z )} {x/f(a)} W={P(f(a), f(f(a)), f(y)), P(f(a), f(z) , z )} {x/f(a), z/f(a)} W={P(f(a), f(f(a)), f(y)), P(f(a), f(f(a)), f(a))} {x/f(a), z/f(a), y/a} W={P(f(a), f(f(a)), f(a)} (v) W={P(x , f(x) , f(a)), P(f(z), f(z), z)} {x/f(z)} w={P(f(z), f(f(z)), f(a)), P(f(z), f(z), z)} next step: occurs check fails: z occurs in f(z) so, NOT UNIFIABLE ------------------------------------------------------------------------ 3b) Using obvious abbreviations and natural-deduction notation, omitting citing "resolution" as rule of inference... also: other solutions are possible! 1. Nab : assumption 2. Nbc : assumption 3. Ga : assumption 4. -Gc : assumption 5. [-Gx, Gy, -Nxy] : assumption 6. [-Ga, Gb] : 1,5 {x/a, y/b} 7. Gb : 6,3 8. [-Gb, Gc] : 5,2 {x/b, y/c} 9. -Gb : 8,4 10. [] : 7,9 ------------------------------------------------------------------------ TENTATIVE GRADING SCHEME FOR PROJECT 1 ------------------------------------------------------------------------ Report + Sample Runs: --------------------- (Note: Please put your sample runs in the body of your report, rather than in the appendix with the code. It might help if you think of the *audience* for your report as someone you are tutoring in logic for AI.): POINTS (*) (*) Point possibilities are of the form "0 n 2n 3n", where 0 means that the work is missing, n points is for an unsatisfactory response, 3n points is for a satisfactory response, and 2n points is for partial credit. ------------------------------------------------------------------------- (1) (a) Description of Clause Form and of your algorithm 0 5 10 15 (b) Correct answers to (b) (by hand) 0 5 10 15 Annotated sample run for (b) 0 5 10 15 Total = 45 points ------------------------------------------------------------------------- (2) (a) Description of unification and of your algorithm 0 5 10 15 (b) Correct answers to (i)-(v) (by hand) 0 1 2 3 for each; total = 15 Annotated sample runs for (i)-(v) 0 1 2 3 for each; total = 15 Total = 45 points ------------------------------------------------------------------------- (3) (a) Description of automated theorem proving using resolution, unification, and refutation 0 5 10 15 (b) Solution (by hand) of the example 0 5 10 15 Annotated sample run for (b) 0 5 10 15 Total = 45 points Grand subtotal = 135 points ------------------------------------------------------------------------- Appendix Containing Documented Code: ----------------------------------- clause form: code: 0 5 10 15 documentation: 0 5 10 15 unification algorithm: code: 0 5 10 15 documentation: 0 5 10 15 resolution: code: 0 5 10 15 documentation: 0 5 10 15 Grand subtotal = 90 points ------------------------------------------------------------------------- Optional bug report: 0 5 10 15 Let's assume that there are parts of your project that you were unable to get to work. If you write an analysis of the problems you encountered and suggest some debugging strategies that you would have tried had I given you another couple of days to work on it, you can earn up to 15 points extra credit. ------------------------------------------------------------------------- Grand total points = 225 472 both 572 --- ---- --- A 214-225 A- 201-213 B+ 189-200 B 176-188 B- 164-175 C+ 151-163 C 126-150 76-150 C- 101-125 D+ 76-100 D 39-75 F 0-38 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun May 1 17:07:43 2005 for ; Sun, 1 May 2005 17:07:43 -0400 (EDT) for ; Sun, 1 May 2005 17:07:42 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 1 May 2005 16:32:39 -0400 20:32:39 -0000 j41KWdjd002842 for ; Sun, 1 May 2005 16:32:39 -0400 (EDT) 16:32:39 -0400 (EDT) Date: Sun, 1 May 2005 16:32:39 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Quiz #2 Grading To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Both sections received SNePS translations of the following form X believes that all Y's are Z, but this isn't the case. I specified in both recitations that you didn't have to use the lex or object-propername caseframe (this was to make taking the quiz and grading the quiz a bit easier)...please *don't* ignore these caseframes on the final!!! Here is the grading schema: Agent-Act-Object (or, better Agent-Act-Action-Object) case frame 0 = missing 3 = incorrectly used OR used different case-frame but didn't give syn/sem 6 = partial credit (mislabeled arc, left unasserted) 9 = correct Rule (forall-ant-cq) case frame 0 = missing 3 = didn't use a rule for "all Y's are Z" AND didn't give syn/sem for new case frame 6-8 = partial credit (asserted, bad antecedent or consequent, used different case frame for "all Y's are Z" other than forall-ant-cq) 9 = correct AndOr connective for "not the case" 0 = missing 1 = incorrectly used 2 = correctly positioned (pointing to proposition for the rule), syntactically incorrect 3 = correct Grade CSE463 Both CSE563 ----------------------------------- A 21 A- 20 B+ 19 B 18 B- 17 C+ 16 C 14-15 10-15 C- 12-13 D+ 10-11 D 6-9 F 0-5 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun May 1 18:34:31 2005 for ; Sun, 1 May 2005 18:34:30 -0400 (EDT) for ; Sun, 1 May 2005 18:34:30 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 1 May 2005 18:13:30 -0400 22:13:30 -0000 for ; Sun, 1 May 2005 18:13:29 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 1 May 2005 18:13:29 -0400 (EDT) Date: Sun, 1 May 2005 18:13:29 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: Albert's comment about the final exam To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: Albert's comment about the final exam ------------------------------------------------------------------------ Albert wrote: | I specified in both recitations that you didn't have to use the lex | or object-propername caseframe (this was to make taking the quiz and | grading the quiz a bit easier)...please *don't* ignore these | caseframes on the final!!! Unless, of course, the instructions on the final exam tell you to ignore them :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sun May 1 18:35:19 2005 for ; Sun, 1 May 2005 18:35:19 -0400 (EDT) for ; Sun, 1 May 2005 18:35:18 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sun, 1 May 2005 18:06:42 -0400 22:06:42 -0000 j41M6gjd006672 for ; Sun, 1 May 2005 18:06:42 -0400 (EDT) 18:06:42 -0400 (EDT) Date: Sun, 1 May 2005 18:06:42 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Programming Project #1: Grading of "description" sections To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello all, I am finished grading your automated theorem provers and, for the most part, they were quite good! The "description" component of the grading rubric might seem like the most subjective...so here is a description of how I graded this component (this is mostly to describe what received a partial credit 10 score): Part I: Clause Form ------------------- a) 0 missing 5 a description which is seriously incorrect 10 a correct description that is not descriptive/too minimal OR a more-than-slighlty incorrect description 15 a correct description I also assigned 10 to students who did not describe their version of FOL for the project (i.e., what symbols were used for wedge, vee, horseshoe etc) anywhere in their paper....the paper should be self-contained...putting this info in a README file is OK, but then you have to cite the README file in your paper! (this was the cruelest I got :) However, I overlooked a missing FOL language description if the notation was that of the newsgroup postings (A, E for quantifiers, > for horseshoe etc.) AND the description included one of the following *nice* features : * Mention that the CF algo is purely syntactic...thus, a computer doesn't need to know the semantics of material conditional (for example) * A detailed explanation of Skolemization * A statement that CF/CNF is a canonical form with limited connectives...thus, we can design an ATP that uses a single rule of inference Part II: Unification --------------------- a) 0 missing 5 seriously incorrect 10 more-than-slighlty incorrect OR correct but no discussion of *any* of: OCCURS-CHECK, mgu, substitution composition, failure conditions, parallel substitution 15 a correct description I gave some 10s for describing unification as taking two wffs as input (in some descriptions you can say that you are "unifying" two clauses and that unification works on predicates in those clauses) , rather than correctly saying that unification works on an atomic and its negation (i.e., literals). Part III: Resolution ------------- a) 0 missing 5 seriously incorrect 10 more-than-slightly incorrect OR correct but missing the logical rule of resolution AND missing a descrip of refutation strategy (can still get a 15 if only one is missing) 15 a correct description I was surprised by how many people didn't even mention Robinson and the rule of resolution...here is a guy who develops a rule of inference to make this whole project possible and most of you didn't even doff your proverbial caps :) ------------------------------------------------------------------------------------------------- For the rest of the grading, I followed the general scheme in Dr. Rapaport's "How I grade" guide http://www.cse.buffalo.edu/~rapaport/howigrade.html namely, 0 = missing n = clearly incorrect 2n = partial credit 3n = correct The only exception was the code/documentation parts: If you didn't have any source code at all for a given section I gave a 0 for the code component and a 15 for the documentation component (no double jeopardy) provided you mentioned somewhere in your report that you could not complete that component. ------------------------------------------------------------------------------------------------- I did not take off any points if you didn't include an abstract...but, if you didn't, you should in the future! An abstract is the brief advertisement that will determine whether or not someone is going to read your paper at all! See you tomorrow, Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Mon May 2 18:45:39 2005 for ; Mon, 2 May 2005 18:45:39 -0400 (EDT) for ; Mon, 2 May 2005 18:45:38 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Mon, 2 May 2005 18:42:12 -0400 22:42:12 -0000 j42MgCjd025935 for ; Mon, 2 May 2005 18:42:12 -0400 (EDT) 18:42:11 -0400 (EDT) Date: Mon, 2 May 2005 18:42:11 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Office Hour Tomorrow To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, Because Dr. Rapaport is not going to be in town, I will hold my office hour tomorrow 1:30-2:30 (as usual). If you are not free then, please send any questions via email (or email me a time you would like to meet on Wed). Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed May 4 17:49:49 2005 for ; Wed, 4 May 2005 17:49:49 -0400 (EDT) for ; Wed, 4 May 2005 17:49:49 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 4 May 2005 17:37:44 -0400 j44LTUjd004544 for ; Wed, 4 May 2005 17:29:30 -0400 (EDT) 17:29:30 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Wed, 4 May 2005 17:29:30 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: SNePS representations by decomposition To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hello All, I got a few questions today about techniques I use for representing complex English sentences in SNePS. I think what I do most frequently is to decompose the complex sentence into its simpler propositions. Remember that a SNePS network is really just a conjunction of all of its asserted nodes. So if you have to represent a sentence like: Albert hopes that all students pass 563. then you can decompose it into several simpler propositions: (1) There is someone (or something :) named Albert (2) Albert hopes (3) (3) All students pass 563. ;;;This is best expressed as a rule Ax[Student(x) > Passes(x,563)] (3antecedent) x is a student (3consequent) x passes 563 Now you must make a decision as to what case frames are available. If you feel that the "standard" case frames are not good enough, then you would invent a new case frame and give your own SYN/SEM (NB: on the exam, YOU WILL BE GIVEN A MENU OF CASE FRAMES TO USE). For this example, the standard case frames will do nicely. So, (sorry, I am pressed for time...I will just give SNePSUL...on the final you will only be responsible for drawing the networks), we represent each of (1),(2), and (3) above ;;;(1) (assert (object B1 propername Albert)) ;;;(2) and (3) (assert (agent B1 act (build action hope object (build forall $x ant (build member *x class student) cq (build agent *x act (build action pass object 563)))))) You should be able to "see" all of the propositional pieces of "Albert hopes all students pass 563" in this representation. hope this technique helps Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Wed May 4 23:08:58 2005 for ; Wed, 4 May 2005 23:08:58 -0400 (EDT) for ; Wed, 4 May 2005 23:08:58 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Wed, 4 May 2005 23:08:33 -0400 j4538Xjd020464 for ; Wed, 4 May 2005 23:08:33 -0400 (EDT) 23:08:33 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Wed, 4 May 2005 23:08:32 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: representing imaginary worlds :) To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU Hi All, A few of you were wondering why I haven't returned your HW #6. HW #6, you will remember, was a "virtual" homework. So if you "virtually" handed it in to your "virtual" TA, please see him to get it back :) I will bring all uncollected work to the final exam tomorrow and you can pick it up. Also, I will post the formula I will be using to compute your recitation grade soon...(it will basically be your best N grades from hw's and quizzes (lowest hw dropped)...where N is to be named later by me and Dr. Rapaport. If you fall really close to the border between, let's say an A- and A, then I will bump you up if you missed no more than M recitations...where M is to be named later by me :) Get some sleep already! Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu May 5 10:06:57 2005 for ; Thu, 5 May 2005 10:06:55 -0400 (EDT) for ; Thu, 5 May 2005 10:06:55 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 5 May 2005 10:06:22 -0400 j45E6Lso025605; Thu, 5 May 2005 10:06:21 -0400 (EDT) 10:06:21 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Thu, 5 May 2005 10:06:21 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: Re: SNePS representations by decomposition Comments: To: ag33@cse.Buffalo.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU | Date: Wed, 4 May 2005 17:29:30 -0400 | From: Albert Goldfain | Subject: SNePS representations by decomposition | | I think what I do most | frequently is to decompose the complex sentence into its simpler | propositions. And what *I* do is then represent those in FOL first, then "translate" to SNePS. (That's one reason I taught FOL before SNePS :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu May 5 10:14:42 2005 for ; Thu, 5 May 2005 10:14:41 -0400 (EDT) for ; Thu, 5 May 2005 10:14:41 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 5 May 2005 10:00:34 -0400 j45E0Xso025424; Thu, 5 May 2005 10:00:33 -0400 (EDT) 10:00:33 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Thu, 5 May 2005 10:00:33 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: HW 10 SOLUTIONS Comments: To: ams67@BUFFALO.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: HW 10 SOLUTIONS ------------------------------------------------------------------------ 1) This representation makes it difficult to easily represent the negation of "Tweety is a bird" or the proposition that someone believes that Tweety is a bird. Points: 0 = no answer 1 = bad/incorrect answer 2 = neither good nor bad ("partial" credit) 3 = good answer ------------------------------------------------------------------------- 2) a) * (a) has a proposition node, (b) doesn't * (b) explicitly represents "bird", (a) doesn't explicitly represent "man" Points: 0,1,2,3 ------------------------------------------------------------------------ b) (a) is legal SNePS; it can be created by: (assert is-a-man John) but (b) is not, because there is no way to build it using SNePSUL Points: 0,1,2,3 ------------------------------------------------------------------------ c) (a) is bad because it doesn't explicitly represent "man" Points: 0,1,2,3 ------------------------------------------------------------------------- d) better representations: (assert object John isa man) or: (assert member John class man), etc. These are better because they fix the problem. Points: 0123 Total points = 12 ------------------------------------------------------------------------- 3) Represents (M1 xor M2); i.e., exactly 1 of {M1,M2} is true Points: 0123 ------------------------------------------------------------------------- 4) Represents that *none* of M1...Mn are true; i.e., -M1 ^ -M2 ^ ... ^ -Mn or: -(M1 v M2 v ... v Mn) Points: 0123 ------------------------------------------------------------------------ 5) Below, I omit "propername" and "lex" nodes for convenience 1. Mary believes everything that Bill believes. SNePSUL code: (assert forall $belief ant (build agent Bill act (build action believe object *belief)) cq (build agent Mary act (build action believe object *belief))) resulting nodes: (m1! (forall v1) (ant (p2 (act (p1 (action believe) (object v1))) (agent Bill))) (cq (p3 (act (p1)) (agent Mary)))) Note that Mary's "act" is the same as Bill's "act"! 0,1,2,3 points ------------------------------------------------------------------------ 2. Mary believes nothing that Bill believes. SNePSUL: (assert forall $belief ant (build agent Bill act (build action believe object *belief)) cq (build min 0 max 0 arg (build agent Mary act (build action believe object *belief)))) nodes: (m1! (forall v1) (ant (p2 (act (p1 (action believe) (object v1))) (agent Bill))) (cq (p4 (min 0) (max 0) (arg (p3 (act (p1)) (agent Mary)))))) 0,1,2,3 points ------------------------------------------------------------------------ 3. It is not the case that Mary believes everything that Bill believes. SNePSUL: (assert min 0 max 0 arg (build forall $belief ant (build agent Bill act (build action believe object *belief)) cq (build agent Mary act (build action believe object *belief)))) nodes: (m2! (min 0) (max 0) (arg (m1 (forall v1) (ant (p2 (act (p1 (action believe) (object v1))) (agent Bill))) (cq (p3 (act (p1)) (agent Mary)))))) 0,1,2,3 points Total = 9 points ------------------------------------------------------------------------- Grand total points = 30 Letter 463 both 563 ------ --- ---- --- A 29-30 A- 28 B+ 26-27 B 24-25 B- 23 C+ 21-22 C 18-20 11-20 C- 14-17 D+ 11-13 D 6-10 F 0-5 From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Thu May 5 21:49:29 2005 for ; Thu, 5 May 2005 21:49:28 -0400 (EDT) for ; Thu, 5 May 2005 21:49:28 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Thu, 5 May 2005 21:11:17 -0400 for ; Thu, 5 May 2005 21:11:16 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Thu, 5 May 2005 21:11:16 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Thu, 5 May 2005 21:11:16 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CORRECTION TO HW 9 ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: CORRECTION TO HW 9 ANSWERS ------------------------------------------------------------------------ A student writes: | In studying for the exam, I noticed a slight mistake in Homework 9's | solutions. | In question 2b, the negation snuck by the forall quantifier without | changing it to and existential quantifier | | > ---> Ey[-Az[-Horse(z) v -Head(y,z)] ^ -Ew[Animal(w) ^ Head(y,w)]] | > ---> Ey[Az.-[-Horse(z) v -Head(y,z)] ^ Aw.-[Animal(w) ^ Head(y,w)]] Oops; sorry. Thanks for pointing this out. Fortunately, it only just barely changes the answer. Here is the corrected answer: ------------------------------------------------------------------------ 2. Answers may vary, depending on how the conclusion is translated. (E.g., another translation is: AyAz[(Horse(z) ^ Head(y,z)) > Ew[Animal(w) ^ Head(y,w)]] ) a) Ax[Horse(x) > Animal(x)] |- Ay[Ez[Horse(z) ^ Head(y,z)] > Ew[Animal(w) ^ Head(y,w)]] (see: http://www.cse.buffalo.edu/~rapaport/Papers/logic,predicate.pdf p. 870) premise: 0123 conclusion: 0123 total = 6 points ------------------------------------------------------------------------ b) CF(premise) = [-Horse(x), Animal(x)] CF(-conclusion): -Ay[Ez[Horse(z) ^ Head(y,z)] > Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[-Ez[Horse(z) ^ Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[Az.-[Horse(z) ^ Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> -Ay[Az[-Horse(z) v -Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> Ey.-[Az[-Horse(z) v -Head(y,z)] v Ew[Animal(w) ^ Head(y,w)]] ---> Ey[-Az[-Horse(z) v -Head(y,z)] ^ -Ew[Animal(w) ^ Head(y,w)]] Above is the old answer, which is correct up to this point. Below is the corrected rest of the answer to this question and to part (c): ---> Ey[Ez.-[-Horse(z) v -Head(y,z)] ^ Aw.-[Animal(w) ^ Head(y,w)]] ---> Ey[Ez[Horse(z) ^ Head(y,z)] ^ Aw[-Animal(w) v -Head(y,w)]] ---> EyEzAw[(Horse(z) ^ Head(y,z)) ^ (-Animal(w) v -Head(y,w))] ---> EzAw[(Horse(z) ^ Head(c,z)) ^ (-Animal(w) v -Head(c,w))], where c is a Skolem constant Up to here, the old answer differs from the new answer only in having an "Ez" where the old one had "Az". Now when we Skolemize, there'll be a further difference. To Skolemize Az.alpha, just drop the quantifier, but to Skolemize Ez.alpha, we need a new Skolem constant; call it d: ---> Aw[(Horse(d) ^ Head(c,d)) ^ (-Animal(w) v -Head(c,w))], ---> Horse(d) ^ Head(c,d) ^ (-Animal(w) v -Head(c,w)) ---> {[Horse(d)], [Head(c,d)], [-Animal(w), -Head(c,w)]} ------------------------------------------------------------------------ Now the resolution proof changes very slightly: c) {[-Horse(x), Animal(x)], (1) [Horse(d)], (2) [Head(c,d)], (3) [-Animal(w), -Head(c,w)]} (4) resolve 1,2 with MGU = {x/d} to get [Animal(d)] (5) resolve 5,4 with MGU = {w/d} to get [-Head(c,d)] (6) resolve 6,3 to get [] From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri May 6 16:35:41 2005 for ; Fri, 6 May 2005 16:35:41 -0400 (EDT) for ; Fri, 6 May 2005 16:35:40 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 6 May 2005 16:35:18 -0400 j46KY4Ym007041; Fri, 6 May 2005 16:34:04 -0400 (EDT) 16:34:04 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Fri, 6 May 2005 16:34:04 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: FINAL EXAM ANSWERS Comments: cc: shapiro@cse.Buffalo.EDU To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: FINAL EXAM ANSWERS ------------------------------------------------------------------------ I have posted the answers to the final exam. Go to the Directory of Documents: http://www.cse.buffalo.edu/~rapaport/563S05/directory.html As soon as we have finished grading, I will post that information to this Listserv so that you can stop by to pick up your exams, projects, etc. From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Fri May 6 22:08:55 2005 for ; Fri, 6 May 2005 22:08:54 -0400 (EDT) for ; Fri, 6 May 2005 22:08:53 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Fri, 6 May 2005 21:48:04 -0400 for ; Fri, 6 May 2005 21:48:03 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Fri, 6 May 2005 21:48:03 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Fri, 6 May 2005 21:48:03 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: CORRECTED FINAL EXAM ANSWERS To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: CORRECTED FINAL EXAM ANSWERS ------------------------------------------------------------------------ Two students wrote: 1. | Question 10 says that M(c) isnt needed, but in fact it is. Mainly | rule 6 which uses 5 (M(c)) and 2. and: 2. | It looks like some of the notation goes wonky starting at problem 7. | I think some of the universal and existential symbols are reversed or | omitted. That or I got some things terribly wrong... 1. M(c) is, indeed, needed. I misinterpreted some notes I was making while I wrote that. I've fixed that remark. 2. I produced the original version of the .pdf file by latexing the original .tex source, converting that to postscript, and then using ps2pdf to create the PDF file. Something got lost in translation. So I deleted the figures (Ziggy and the SNePS syntax and semantics) and used pdflatex. The new file should be a bit more readable. Please let me know if there are any more errors. Note that none of the above affect your grades :-) From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Sat May 7 11:01:47 2005 for ; Sat, 7 May 2005 11:01:47 -0400 (EDT) for ; Sat, 7 May 2005 11:01:47 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Sat, 7 May 2005 10:45:26 -0400 for ; Sat, 7 May 2005 10:45:20 -0400 (EDT) -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Sat, 7 May 2005 10:45:20 -0400 Reply-To: Albert Goldfain From: Albert Goldfain Subject: Formula for computing Recitation Grades To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU What follows is the algorithm I used to compute your Recitation component for your grade. First a summary of the work: Check/Check+: HW1,HW2,MyExtraHW(optional) Graded: HW3,HW4,HW5,HW7,HW8,HW9,Quiz1,Quiz2,HW10(optional) And here is the algorithm: 1. Drop the lowest of HW3,HW4,HW5,HW7,HW8,HW9 2. Add a half letter grade (e.g., A- to A) to lowest remaining individual "Graded" HW for every Check+ earned on HW1,HW2,MyExtraHW. Call this resulting set X. 3. Take the best 6 members from X union {Quiz1,Quiz2,HW10} 4. Average the 6 grades on the university 4 pt scale (A=4.0,A-=3.7, etc.) 5. If the average lies .05 or less from changing a letter grade, I will bump you up if you missed less than 4 recitations (i.e., if you missed 3, 2 or 1 recitations) Albert From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue May 10 13:43:58 2005 for ; Tue, 10 May 2005 13:43:58 -0400 (EDT) for ; Tue, 10 May 2005 13:43:58 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 10 May 2005 13:42:25 -0400 j4AHgOYm001204 for ; Tue, 10 May 2005 13:42:24 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Tue, 10 May 2005 13:42:24 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Tue, 10 May 2005 13:42:24 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: CSE 4/563: FINAL EXAM AND COURSE GRADES To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU ------------------------------------------------------------------------ Subject: CSE 4/563: FINAL EXAM AND COURSE GRADES ------------------------------------------------------------------------ Grading is done (almost)! (I'm just waiting for one student to turn in a late paper or to tell me that he's not going to.) ======================================================================== PLEASE STOP BY MY OFFICE TO PICK UP YOUR FINAL EXAM, YOUR PROJECT 2, AND ANY OTHER LEFTOVER PAPERS I MIGHT HAVE (INCLUDING SOME FROM LAST YEAR'S CSE 410/510 :-) ======================================================================== !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! (If you don't, I'll have to keep them in my office for the next 3 years:-| !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Some statistics: CSE 463: Final Exam: Range: 56-136/136 points Average: 116 points; B+ Course Grade: Range: D to A Average: B+ CSE 563: Final Exam: Range: 95-136/136 points Average: 124 points; A- Course Grade: Range: B+ to A Average: A- I have posted the grades for CSE 463, so you should be able to find out your grade from MyUB, probably tomorrow. I will soon post the grades for CSE 563. In the meantime, please feel free to email me to find out your own grades. Finally, I enjoyed teaching the course; I hope you enjoyed taking it. I intend to analyze your answers to the Logical Illusion problem at the end of the exam, and will post the results to the Listserv. -Bill Rapaport ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html From owner-cse563-sp05-list@LISTSERV.BUFFALO.EDU Tue May 10 14:17:13 2005 for ; Tue, 10 May 2005 14:17:13 -0400 (EDT) for ; Tue, 10 May 2005 14:17:11 -0400 (EDT) CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU; Tue, 10 May 2005 14:16:47 -0400 j4AIGjYm001616 for ; Tue, 10 May 2005 14:16:45 -0400 (EDT) cse563-sp05-list@listserv.buffalo.edu; Tue, 10 May 2005 14:16:45 -0400 (EDT) X-PM-EL-Spam-Prob: : 7% Date: Tue, 10 May 2005 14:16:45 -0400 Reply-To: "William J. Rapaport" From: "William J. Rapaport" Subject: office hours during summer To: CSE563-SP05-LIST@LISTSERV.BUFFALO.EDU For those of you wanting to stop by to get your exams, etc.: I should be in most of Wednesday, May 11, but email/call first to make sure. Ditto Friday, May 13. I'll be in and out throughout the summer, and will be around in the Fall. ------------------------------------------------------------------------ William J. Rapaport Associate Professor of Computer Science/Adjunct Professor of Philosophy Member, Center for Cognitive Science Associate Director, SNePS Research Group (SNeRG) 201 Bell Hall (office: 214 Bell) | 716-645-3180 x 112 Dept. of Computer Science & Engineering | fax: 716-645-3464 University at Buffalo (SUNY) | rapaport@cse.buffalo.edu Buffalo, NY 14260-2000 | ------------------------------------------------------------------------ CSE: www.cse.buffalo.edu/ homepage: www.cse.buffalo.edu/~rapaport/ SNeRG: www.cse.buffalo.edu/sneps/ Buffalo Restaurant Guide: www.cse.buffalo.edu/restaurant.guide/ Cognitive Science: wings.buffalo.edu/cogsci/ Good Things about Buffalo: www.cse.buffalo.edu/~rapaport/buffalo.html