CS 528 - Graphics/Geometry/Vision/Robotics Seminar

Course announcement

Marc Levoy
Pat Hanrahan
Leo Guibas
Carlo Tomasi
Oussama Khatib
Jean-Claude LaTombe
Gene Golub

To all students in graphics, geometry, vision, robotics, and numerical methods:

A few weeks ago, your advisors (Marc Levoy, Pat Hanrahan, Leo Guibas, Carlo Tomasi, Oussama Khatib, Jean-Claude LaTombe, and Gene Golub) got together to chat about ways to foster more interaction between our research groups. Many of you already straddle the boundary between two or more of these areas, and are doing some very exciting research; we'd like to see more of this. One of the ideas we came up with is an every-other-week seminar series that brings together all five groups. Here is the description of this seminar series as it will appear in this year's course catalogue:

CS 528: Graphics/Geometry/Vision/Robotics Seminar - Weekly series of informal research talks on topics related to perceiving, modeling, manipulating, and displaying the physical world. Also of interest are the computational models and numerical methods underlying these topics. This seminar was created to bring together the faculty and students in these five closely related areas. 1 unit, Aut, Win, Spr (Staff).

Our goal for this seminar series is 100% attendance by all faculty and students in the five areas. Students may register for CS 528 (for 1 credit) if they like, but it is not mandatory. Talks will be in B12 (the slant-floored lecture room in the basement) starting at 4:15pm every other Wednesday.

To get the ball rolling, our first meeting will be a reception! We profs will give an overview of our motivations for combining forces, the philosophy behind the physiqual, and our reasons for organizing this seminar series. We'll then go around the room giving each student a chance to introduce himself or herself and briefly state their area of research. We'll end with some kind of entertainment (maybe an excerpt from the Siggraph film/video show?) Refreshments will be served.

After this icebreaker, meetings will be biweekly. We will rotate in an orderly fashion among the five groups: 5 areas x biweekly = 1 talk per area per quarter. In the autumn quarter, the professors will speak, one every other week. In the winter and spring quarters, each group will be responsible for designating a student to speak. Remember - the purpose for these talks is to instruct and excite people in other disciplines. Thus, talks should include enough background material so that non-specialists can follow it, and should include plenty of discussion, speculation, enumeration of possible applications in other areas, etc. Conference talks are therefore not appropriate. More on this later.

To summarize, the schedule for the autumn quarter is as follows:

	Wed., Oct.  1	Reception, overview, and introductions  (everybody!)
	Wed., Oct. 15	Talk by graphics group		(Levoy or Hanrahan)
	Wed., Oct. 29	Talk by geometry group		(Guibas)
	Wed., Nov. 12	Talk by vision group		(Tomasi)
	Wed., Nov. 19	Talk by robotics group		(Khatib or LaTombe)
	Wed., Dec.  3	Talk by numerical methods group	(Golub)

(Note the hiccup in the biweekly rhythm just before Thanksgiving.)

So, please set aside these Wednesday afternoons at 4:15, and plan to be there at the October 1 reception, which will be in B12 unless we announce otherwise. We profs firmly believe that this unusual form of cross-fertilization will generate lots of new ideas, not to mention some great dissertation topics!

September 10, 1997