November 28, Monday 12:10, Room 665, Education Building
** NOTE SPECIAL TIME AND PLACE **

Title: Computing, Learning, and Classifying, Under the Inverse

Lecturer: Tsvi Achler

Lecturer homepage : http://reason.cs.uiuc.edu/tsvi/

Affiliation : Los Alamos National Labs

 

The brain has recognition capabilities that remain unmatched bycomputer algorithms. We hypothesize that using matchedfeedforward-feedback connections, recognition centers of thebrain reconstruct an internal copy of inputs using knowledge the brainhas previously accumulated, in accordance with a class models called"generative models". Subsequently, it minimizes the error between theinternal copy and the input from the environment. We study how thisstrategy may enable simple flexible learning, scalability, overcomecombinatorial problems associated with pattern mixtures, anddisplay cognitive phenomena.