Dharmendra S. Modha

My Work and Thoughts.

Archives for November 2006

“What we know and don’t know about biological vision”

By

Today, we had an incredible time with Professor Bruno Olshausen who heads the Redwood Center at UC Berkeley, and is Associate Professor, Helen Wills Neuroscience Institute and School of Optometry, UC Berkeley.

He described the state-of-the-art in the knowledge about the primary visual cortex (V1). Here is the abstract:

Nervous systems have evolved impressive abilities to extract useful information about the environment from images. This talk will review some highlights of what has been learned from a combination of psychophysical, neurophysiological, and computational studies of the visual system, in addition to those aspects that remain a mystery. I will also describe some new organizing principles which provide hope for elucidating the function of visual cortex.

TOP500 Supercomputing Sites

By

At Supercomputing 06 conference, the 28th TOP 500 Supercomputing Sites List was released.

Please see the original article at: http://top500.org/lists/2006/11

"IBM BlueGene/L system, installed at DOE’s Lawrence Livermore National Laboratory (LLNL), retains the No. 1 spot with a Linpack performance of 280.6 teraflops (trillions of calculations per second, or Tflop/s)".

The BlueGene/L at IBM’s Almaden Research Center ranked 64th on the list with a Linpack performance of 9.43 teraflops, for details, please see: http://top500.org/list/2006/11/100 

Dr. Gerry Tesauro

By

Dr. Gerry Tesauro is a Research Staff Member in computer science at the IBM TJ Watson Research Center in Hawthorne, NY. He is famous for developing TD-Gammon a self-teaching program that learned to play backgammon at human world championship level. TD-Gammon uses reinforcement learning — a well known technique in Machine Learning that has a close connection to the role of dopamine neurons in the brain.

Gerry was invited to Almaden by Nimrod Megiddo, and gave us a wonderful talk on how he is applying reinforcement learning to improving system management policies.

References:

1. Gerald Tesauro, Nicholas K. Jong, Rajarshi Das, Mohamed N. Bennani: Improvement of Systems Management Policies Using Hybrid Reinforcement Learning. ECML 2006: 783-791

2. Gerald Tesauro: Online Resource Allocation Using Decompositional Reinforcement Learning. AAAI 2005: 886-891

3. Gerald Tesauro: Temporal Difference Learning and TD-Gammon. Commun. ACM 38(3): 58-68 (1995)