Current Graduate Course Descriptions (Spring 2009)
Broadly speaking, we may distinguish two types of cyber physical environments: indoor and outdoor. We will focus on indoor environments, a topic that is of particular interest for our department as we plan the 'smart room' of the new building, which is expected to be ready for occupancy by 2011. Indoor environments do not incur some of the challenges in power-supply and bandwidth that confront outdoor environments.
We are interested in biometric-based cyber physical systems, in which occupants are identified and tracked using biometric capture devices such as cameras, microphones, etc. We are also interested in answering spatio-temporal queries about the whereabouts of occupants. Such systems can be used in homes for the elderly or disabled, department stores, and eventually in larger arenas such as airports.
In this seminar we will explore the role of declarative knowledge and reasoning as a means of enhancing the recognition capabilities of biometric-based cyber physical systems. Towards this end, we will explore how to characterize the performance of the environment in terms of the information-theoretic concepts of precision and recall. Some preliminary approaches are described in papers 2008-1 and 2008-12 in CSE Technical Report Archive. We are interested in exploring how recognition and reasoning can be augmented with user interaction via speech and gestures, and actions that effect changes in the environment; as well as environments that improve their performance by learning from past behaviors. Other topics can be included depending upon student interest.
Students taking the seminar will be expected to make one or two class presentations and contribute towards an implementation or work on a research paper, with the approval of the seminar instructor. Students can participate in ongoing research projects in this area, and engage in experimental as well as conceptual/theoretical research.
+ Course Requirements: Class Presentation and Course Project
+ Grading Scheme: S/U, as per Department policy
Please email Dr. Jayaraman (bharat@cse.buffalo.edu) if you are interested in taking the seminar for credit.
PREREQUISITES:Graduate Standing, Interest in Artificial Intelligence
WEB PAGE: http://www.cse.buffalo.edu/LRG/CSE734/