[Colloquium] Bottou talk - Mon. 10/27, 2:30 at TTI

[Meridel Trimble]

*TOYOTA TECHNOLOGICAL INSTITUTE TALK*
Time: 2:30pm (Not 12:15 as previously scheduled)
Date: October 27th, 2003

Speaker: Léon Bottou
NEC Research Institute
Speaker's homepage http://leon.bottou.org/

Title: Stochastic Gradient Learning Algorithms

A general framework for stochastic learning algorithms is first presented.  
This framework encompasses the most common online learning algorithms in use 
today, as illustrated by several examples.  The stochastic approximation theory 
then provides general results describing the convergence of all these learning 
algorithms at once.  Finally we consider situations where training data is 
abundant and computing resources are comparatively scarce. We argue that 
suitably designed stochastic learning algorithms asymptotically outperform 
batch learning algorithm. Both theoretical and experimental evidences are 
presented. 

For further information, please contact Meridel (mtrimble at tti-c.org/834.9873)

*REFRESHMENTS PROVIDED*