Interdisciplinary Ph.D. Program in Cognitive Science
UNIVERSITY OF CALIFORNIA, SAN DIEGO


Cognitive Science 200: Visual salience: neurophysiology and models
(Section ID: 659727)
Friday afternoons, CSB 003
Student discussion session: 2-2:50PM
Public lecture: 3-4:50
Organizers: Gary Cottrell, Javier Movellan, and Mike Mozer

Fall 2009


To join the cs200 mailing list to receive announcements of talks, see this instruction page. Cognitive Science 200 is an interdisciplinary seminar of changing topics, and is used as a mechanism for Ph.D. students in the Interdisciplinary Ph.D. Program and in the Cognitive Science Department to achieve breadth. This quarter, the topic is "Visual Salience: neurophysiology and models."

As a result of having a foveated retina, we actively move our eyes to direct our highest resolution of visual processing towards interesting things. In fact, we move our eyes about three times a second; it is a decision we make about 172,000 times a day, more than any other in our lives! How do we decide where to look? Where we look has two major influences: 1) exogenous or "bottom-up" influences - "busy-ness" of some sort in our visual field, such as motion or areas of high contrast, and 2) endogenous or "top-down" influences, i.e., our current goals or task. This quarter, COGS 200 will continue a conversation/debate that has been going on among several researchers in the Temporal Dynamics of Learning Center (tdlc.ucsd.edu) with somewhat distinct models of visual salience. We have also invited researchers with expertise in the neurophysiology of the representation of "attention" or "salience" in the brain. Students in the course will be required to read papers and to write a paper at the end of the quarter.

The room for Cogsci 200 is Cognitive Science Building 003. The meeting times are Fridays 2-2:50PM for registered students, and 3:00-4:50PM for the lectures (to which the UCSD Cognitive Science community is invited). This will be followed usually by the cognitive science happy hour in the cog sci building courtyard.

The graduate student section from 2-2:50 will involve the professor using the dreaded index card method: students will be asked questions about the papers that are intended to generate some discussion and understanding of the material. Students are therefore expected to have done the reading before class. The method involves index cards with every student's name on them. These are shuffled at the beginning of class, and then students are asked questions in order of their appearance on the card. The first question is almost always, "What is the point of this paper?", and is often asked several times until we converge on one or more main themes of the paper.

The requirements for the class are: 1) reading the assigned papers; 2) being able to answer questions about them in discussion section; 3) asking the speaker a question about 20% of the time (I'll be keeping track! I.e., you need to ask 2 questions all quarter) and 4) writing an approximately 10 page research proposal that is of your own choosing - it could be an extension to one of the topics covered in the lectures, tesing a hypothesis about salience or attention, pitting the various models against one another, etc. It should be specific enough that there are clear criteria for success or failure. The draft of this is due in the 8th week, the final version is due on the Monday of finals week.


DATE PRESENTER TITLE
PAPER
DISCUSSION
PAGE
SLIDES
September 25
Gary Cottrell
Organizational meeting



October 30
Matt Wilder
A Unified Theory of Attentional Control
Matthew H. Wilder, Michael C. Mozer and Christopher D. Wickens (draft) A Unified Theory of Attentional Control Discuss any paper here...
November 6
Randy O’Reilly (CU Boulder)
Attention without an attentional mechanism
Chapter 8 of CCNC (read 8.5 and 8.6)
Cohen et al. 1994
Discuss any paper here...
November 20
Laurent Itti (USC)
Computational modeling of surprise and combined bottom-up/top-down gaze control
Itti, Laurent, and Baldi, Pierre (2009) Bayesian surprise attracts human attention. Vision Research 49:1295–1306
R. J. Peters, L. Itti (2007) Congruence between model and human attention reveals unique signatures of critical visual events, In: Advances in Neural Information Processing Systems, Vol. 20 (NIPS*2007), pp. 1145-1152, Cambridge, MA:MIT Press.
Discuss any paper here...
December 4
Angela Yu (UCSD)
Attentional Selection: Toward a Rational Bayesian Account
Yu, Dayan & Cohen (2009) Dynamics of Attentional Selection Under Conflict: Toward a Rational Bayesian Account [pdf]
Yu & Dayan (2004) Inference, Attention, and Decision in a Bayesian Neural Architecture. In Advances in Neural Information Processing Systems (NIPS04). [pdf]
Discuss any paper here...

The instructor is Professor Gary Cottrell, whose office is CSE Building room 4130. Feel free to send email to arrange an appointment, or telephone (858) 534-6640.

REGISTRATION

Students may take the seminar only for four units of S/U credit. Students should register for COGS 200, section id 659727. If you must have a letter grade (because of your departmental requirements), please see me and let me know!



Most recently updated on September 29, 2009 by Gary Cottrell, gary@ucsd.edu

AltStyle によって変換されたページ (->オリジナル) /