Seminars in December, 1996
- 12/10/96 - Recent work in computational economic, Professor Yoav Shoham, Computer Science Department, Stanford University.
- 12/03/96 - Richard Ivry, University of California, Berkeley.
Seminar Abstracts
Tuesday December 10,1996
4-5:30pm
306 Soda Hall
Recent work in computational economic
Professor Yoav Shoham
Computer Science Department
Stanford University
Abstract:
In recent years I've found myself together with students and other colleagues doing various things that smack of economics. I'll give an overview of some of our recent work, and then cover one or two of the following topics in more detail:1. (Designed) social laws and (emergent) social conventions in a computational environment. Joint with Moshe Tennenholtz.
2. An application to content-based/collaborative recommendation system for the internet. Joint with Marko Balabanovic.
3. Utility distributions, or, how utilities can enjoy the rich structure of probabilities.
December 3,1996
4-5:30pm
306 Soda Hall
Richard Ivry
University of California, Berkeley
Abstract:
The cerebellum can be characterized as an internal timing mechanism. Patients with lesions of the neocerebellum are impaired on various tasks that require precise timing, not only in terms of motor control, but also in perception. Evidence for motor timing problems comes from studies of repetitive tapping and speech production. Evidence for perceptual problems comes from studies of duration discrimination, velocity perception, and perhaps certain forms of speech perception. The timing hypothesis also provides parsimonious account of the role of the cerebellar cortex in a simple form of sensorimotor learning, eyeblink conditioning. Recent work indicates that rather than the cerebellum represents time in a distributed format; that is, as a system of multiple timers rather than a single clock. This hypothesis is consistent with a view in which the timing mechanisms do not involve a pacemaker process, but rather operate as interval-based timers in which separable elements code distinct intervals. Evidence in support of this view comes from research involving bimanual tapping. The cerebellar timing system is one module within a distributed system for the control of coordinated action. The cerebellum does not appear to play a critical role in other aspects of coordination such as spatial planning, temporal coupling, and movement selection.
[Top|Home]