Valentino Crespi, Ph.D. - January 27, 2010
California State University, Los Angeles
Department of Computer Science
"Trackability and Machine Learning of Behaviors and Processes"
In this talk I will start by introducing an original and rigorous concept of "trackability" of processes in a distributed sensing system. The purpose of this notion of trackability is to determine the "complexity" of estimating state trajectories of a target process based on a discrete-time sequence of noisy "observations". Then I will present new algorithms to learn Hidden Markov Models (HMMs) from typical realizations of the associated stochastic process. Those methods are based on the non-negative matrix factorization (NMF) of higher order Markovian statistics and are structurally different from the classical Baum-Welsh and associated approaches.
Lisette De Pillis, Ph.D. - February 10, 2010
Harvey Mudd College
Department of Mathematics
"Mathematical Modeling of Cancer Growth and Immunological Indirection"
Abstract: Immunotherapy, a treatment approach that enhances the body's
natural ability to fight cancers, is becoming increasingly prevalent in
many multi-stage treatment programs that also include chemotherapy,
radiation and surgery. The critical importance of the immune system
in combating cancer has been verified both clinically and through mathematical
models. In this talk, we will discuss the biological and mathematical
sides of the question of how cancer grows, how the cancer interacts
with the immune system, and treatment approaches that harness the
power of the immune system.
Christof Koch, Ph.D. - March 10, 2010
California Institute of Technology
Professor of Cognitive and Behavioral Biology and
Professor of Computation and Neural Systems
ÂThe Science of Consciousness: What do we Know and What do we Need to Know?Â
Abstract: At the beginning of the third millennium, a Science of Consciousness is beginning to emerge that seeks to understand how subjective conscious sensations, qualia, emerge from certain types of organized matter, in particular from the mammalian cortico-thalamic system. I will summarize what is known about the neurobiology of consciousness, outline the limits to our knowledge, and describe ongoing psychophysical and electrophysiological experiments to manipulate the relationship between physical stimuli and their associated conscious percepts and to discover and characterize the neuronal correlates of consciousness (NCC). I will conclude by discussing a very promising theoretical approach to consciousness - Integrated Information Theory by Tononi - grounded in circuit complexity and information theory.
Seminars are from 3:00-4:00 p.m. in Biological Sciences room 244