:::

Announcements

:::

【2011-06-22】Prof. James Ting-Ho Lo (University of Maryland Baltimore County), "A Cortex-Like Learning Machine for Temporal Hierarchical Pattern Clustering, Detection, and Recognition"

非專題討論演講公告
Poster:Post date:2011-05-24

Title: A Cortex-Like Learning Machine for Temporal Hierarchical Pattern Clustering, Detection, and Recognition
Speaker: Prof. James Ting-Ho Lo
, (University of Maryland Baltimore County)
Time: 14:20, June 22 (Wed), 2011
Place:   Room 310, CSIE Building

Abstract:

A new paradigm of machine learning, which is also a Low-Order Model (LOM) of biological neural networks, will be proposed. LOM is a network of biologically plausible models of dendritic nodes/trees, spiking/nonspiking neurons, unsupervised/supervised covariance/accumulation learning mechanisms, feedback connections, and a scheme for maximal generalization.

These component models were motivated and necessitated by making LOM learn and retrieve easily; and cluster, detect and recognize multiple/hierarchical corrupted, distorted and occluded temporal and spatial patterns.

On one hand, with many features and capabilities desirable of a learning machine, LOM is expected to be a powerful engine for intelligent systems. On the other hand, biological plausibility of LOM makes a strong case that LOM is the common cortical algorithm long hypothesized by neuroscientists.

Short-bio:

James Ting-Ho Lo is a Professor in the Department of Mathematics and Statistics. He received the Ph.D. degree from the University of Southern California and was a Postdoctoral Research Associate at Stanford University and Harvard University. His research interests have included optimal filtering, system control and identification, active noise and vibration control, and computational intelligence. In 1992, he solved the long-standing notorious problem of optimal nonlinear filtering in its most general setting and obtained a best paper award.
Subsequently, he conceived and developed adaptive neural networks with long- and short-term memories, accommodative neural network for adaptive processing without online processor adjustment, and robust/adaptive neural networks with a continuous spectrum of robustness; which constitute a systematic general approach to effective robust or/and adaptive processing for system control/identification and signal processing.
He developed the convexification method for avoiding poor local-minima in data fitting (e.g., training neural networks and estimating regression models), removing a main obstacle in the neural network approach and nonlinear regression in statistics.
In recent years, Dr. Lo developed a functional and a low-order model of biological neural networks. The former, called the temporal hierarchical probabilistic associative memory (THPAM), is a new paradigm of learning machines. The latter, the low-order model, comprises biologically plausible models of dendritic nodes/trees, synapses, spiking/nonspiking somas, unsupervised/supervised learning mechanisms, a maximal generalization scheme, and feedbacks with different delay durations; which integrate into a biologically plausible learning/retrieving algorithm and answer numerous fundamental questions in neuroscience.




Last modification time:2011-05-24 PM 3:03

cron web_use_log