Department of Computer Science State University of New York at Buffalo CS 705: SEMINAR: MACHINE TRANSLATION Semester: Fall 1986 Instructor: William J. Rapaport Time: tba Office: Bell 214 Room: tba Phone: 636-3193 Office Hours:tba & by appointment Email: rapaport TOPICS: o+ Machine translation (MT)-the use of a computer to provide automatic translations from one natural language to another-was one of the earliest goals of AI. It may even be said to pre-date AI, since the earliest suggestions for it were suggested by code-breaking strategies developed during World War II. The early work was generally deemed a failure, because it did not take semantics-or ``world knowledge''-into account. Recent work, however, has apparently succeeded far beyone anyone's hopes after that initial failure. As seen from an AI perspective, MT embodies many of the major problems of natural-language understanding: parsing, knowledge representation and reasoning, and generation. Moreover, differences between fully automatic MT and machine-aided-translation (MAT) (either interactive, or with pre- or post-editing) raise interesting questions about computer-human interaction and about the attainability of some of the goals of AI. o+ This seminar will be focused around careful readings of some of the important early writings about MT, as well as inves- tigations of recent MT and MAT systems. I will also present some of the standard natural-language processing techniques for parsing and generating (so that this course may serve as a brief introduction to natural-language processing). MECHANICS OF THE COURSE: o+ Students registered for the seminar will be expected to write a 10-15 page term paper and/or build a small (but interesting) M(A)T system (accompanied by a full report on the system). There will also be an oppportunity for (and perhaps even a requirement of!) seminar presentations. o+ The grade will be a weighted average of seminar participa- tion and the project grade (probably 2/3-1/3). TENTATIVE SCHEDULE (1) Weaver, Warren (1949), ``Translation,'' in W. N. Locke & A. D. Booth (eds.), Machine Translation of Languages (Cam- bridge, MA: MIT Press, 1955): 15-23. (2) A. Bar-Hillel, Yehoshua, ``Some Linguistic Obstacles to Machine Translation,'' Appendix II of Y. Bar-Hillel, ``The Present Status of Automatic Translation of Languages,'' in F. L. Alt (ed.), Advances in Computers (New York: Academic Press, 1960): 91-163. B. Bar-Hillel, Yehoshua, ``A Demonstration of the Nonfeasibil- ity of Fully Automatic, High Quality Translation,'' Appendix III of Y. Bar-Hillel, ``The Present Status of Automatic Translation of Languages,'' in F. L. Alt (ed.), Advances in Computers (New York: Academic Press, 1960): 91-163. (3) The ALPAC Report. (4) Slocum, Jonathan (1985), ``A Survey of Machine Transla- tion: Its History, Current Status, and Future Pros- pects,'' Computational Linguistics 11: 1-17. (5) Issues in machine translation. (6) Natural-language processing: parsing, knowledge represen- tation, generation; SNePS. (7) Seminar presentations on particular MT systems. (8) A. Quine, Willard Van Orman (1960), Word and Object, Chaps. 1-2 (Cambridge, MA: MIT Press). B. Quine, Willard Van Orman (1969), ``Ontological Relativity,'' in W. V. O. Quine, Ontological Relativity and Other Essays (New York: Columbia University Press): 26-68. (9) White, John S.; Kay, Martin; and King, Margaret (1986), ``Forum on Machine Translation,'' Proc. 24th Annual Meet- ing ACL (Columbia Univ.) (Morristown, NJ: Assoc. Comp. Ling.): 267-70.