2010 Distinguished Lecture Series
March 9, 2010
Pat Hanrahan, CANON Professor, Stanford University
Title: Why are Graphics Systems so Fast? (click to see presentation)
Abstract:
Over the last decade graphics hardware has become a key component of mobile and personal computers. Most programmers understand CPUs well, but have a limited understanding of GPUs (Graphics Processing Units). GPUs are viewed as specialized hardware optimized for rendering. That view is not accurate. Instead, they are best characterized as parallel computers that combine many cores, many threads, and wide vector processing units. In this talk, I will describe the architectures of different GPUs built by AMD, NVIDIA and Intel (the new Larrabee processor). I will also discuss the programming models that are used to achieve high performance on such heterogeneous architectures. The innovative combination of processor design and programming model are why graphics systems are so fast.
About the speaker:
Pat Hanrahan is the CANON Professor of Computer Science and Electrical Engineering at Stanford University where he teaches computer graphics. His current research involves visualization, image synthesis, virtual worlds, and graphics systems and architectures. Before joining Stanford he was a faculty member at Princeton. He has also worked at Pixar where he developed volume rendering software and was the chief architect of the RenderMan Interface - a protocol that allows modeling programs to describe scenes to high quality rendering programs. Professor Hanrahan has received three university teaching awards. He has received two Academy Awards for Science and Technology, the Spirit of America Creativity Award, the SIGGRAPH Computer Graphics Achievement Award, the SIGGRAPH Stephen A. Coons Award and the IEEE Visualization Career Award. He was recently elected to the National Academy of Engineering and the American Academy of Arts and Sciences.
May 13, 20120
David P. Williamson, Cornell University
Title: The Rank Aggregation Problem
Abstract:
The rank aggregation problem was introduced by Dwork, Kumar, Naor, and Sivakumar (WWW 2001) in the context of finding good rankings of web pages by drawing on multiple input rankings from various search engines. The work draws on earlier work in the social choice literature on combining voter preferences specified by ranked alternatives. Dwork et al. propose finding a ranking that minimizes the sum of its Kendall distance to the input rankings; this can be viewed as a weighted feedback arc set problem. I will give an overview of the rank aggregation problem and some of its applications, including an application to intranet search (WWW 2003). I will also cover work done on finding good approximation algorithms for the problem. In particular, Ailon, Charikar, and Newman (STOC 2005) introduce a randomized "pivoting" algorithm for rank aggregation and related problems; recent work has extended this to deterministic pivoting algorithms for constrained versions of the problem (van Zuylen, Hegde, Jain and Williamson, SODA 2007, van Zuylen and Williamson 2009) and has yielded a polynomial-time approximation scheme for the problem (Kenyon- Mathieu and Schudy, STOC 2007). Experimental work of Schalekamp and van Zuylen has given us a good sense of the tradeoffs of these various algorithms in practice. If time permits, I will discuss some extensions of this work to finding good clusterings and hierarchical clusterings.
About the speaker:
David Williamson was born in Madison, Wisconsin, but grew up in the suburbs of Honolulu, Hawaii. He received his Ph.D. in 1993 from the Massachusetts Institute of Technology. In 1995 he joined IBM Research, and from 2000-2003 was the Senior Manager of the Computer Science Principles and Methodologies Department at IBM's Almaden Research Center. In 2004, he joined Cornell University as a professor in the School of Operations Research and Information Engineering, and the Faculty of Computing and Information Science.
David is well-known for his work on the topic of approximation algorithms, and is a coauthor of the forthcoming book "The Design of Approximation Algorithms", to be published by Cambridge University Press. His Ph.D. dissertation on designing low-cost survivable networks was awarded several prizes, including the 1996 SIAM DiPrima Prize and the 1994 Tucker Prize from the Mathematical Programming Society. His work with Michel Goemans on the uses of semidefinite programming in approximation algorithms was awarded the 1999 SIAM Activity Group on Optimization prize, and the 2000 Fulkerson Prize from the Mathematical Programming Society and the American Mathematical Society. He serves as an associate editor on several journals, and is the former Area Editor for Discrete Optimization on the journal Mathematics of Operations Research.
June 9, 2010
David Patterson, Pardee Professor of Computer Science, University of California at Berkeley
Title: Revolution Has Started: Are You Part of the Solution or Part of the Problem? (see the presentation notes (PDF))
Abstract:
This talk will explain * Why the La-Z-Boy era of programming is over (Hint: clock rates aren't increasing) * The opportunities and pitfalls of this revolution * What Berkeley is doing to try to be near the forefront of this revolution (A paper describing the efforts of Berkeley's Par Lab appeared on the cover of the October 2009 issue of Communications of the ACM. If time is available, the talk will explain an intuitive visual performance model called "Roofline," which appeared in April 2009 CACM.)
About the speaker:
David Patterson is the Pardee Professor of Computer Science at the University of California at Berkeley, which he joined after graduating from UCLA in 1977. Dave's research style is to identify critical questions for the IT industry and gather inter-disciplinary groups of faculty and graduate students to answer them. The answer is typically embodied in demonstration systems, and these demonstration systems are later mirrored in commercial products. In addition to research impact, these projects train leaders of our field. The best known projects were Reduced Instruction Set Computers (RISC), Redundant Array of Inexpensive Disks (RAID), and Networks of Workstations (NOW). A measure of the success of projects is the list of awards won by Patterson and as his teammates: the C & C Prize, the IEEE von Neumann Medal, the IEEE Johnson Storage Award, the SIGMOD Test of Time award, and the Katayanagi Prize. He was also elected to the American Academy of Arts and Sciences, National Academy of Engineering, National Academy of Sciences, and the Silicon Valley Engineering Hall of Fame. Most recently, he was named a Fellow of the Computer History Museum. The full list includes about 20 awards for research, teaching, and service. In his spare time he coauthored five books, including two with John Hennessy, who is President of Stanford University. Patterson also served as Chair of the Computer Science Division at UC Berkeley, Chair of the Computing Research Association and President of ACM.
October 14, 2010
Pietro Perona, Allen E. Puckett Professor of Electrical Engineering, California Institute of Technology
Title: Vision of a Visipedia
Abstract:
The web is not yet perfect: while text is easily searched and organized, pictures (the vast majority of the bits that one can find online) are not. In order to see how one could make pictures first-class citizens of the web, I explore the idea of Visipedia, a visual interface for Wikipedia that is able to answer visual queries and enables experts to contribute and organize visual knowledge. Five distinct groups of humans would interact through Visipedia: users, experts, editors, visual workers, and machine vision scientists. The latter would gradually build automata able to interpret images. I explore some of the technical challenges involved in making Visipedia happen. I argue that Visipedia will likely grow organically, combining state-of-the-art machine vision with human labor.
Joint work with P. Welinder, S. Belongie, S. Branson, K. Wah
About the speaker:
Dr. Pietro Perona is the Allen E. Puckett Professor of Electrical Engineering at Caltech. He directs Computation and Neural Systems (www.cns.caltech.edu), a PhD pogram centered on the study of biological brains and intelligent machines. Professor Perona's research centers on vision. He has contributed to the theory of partial differential equations for image processing and boundary formation, and to modeling the early visual system's function. He is currently interested in visual categories and visual recognition.