Seminar on Computational Learning and Adaptation


  Epsilon Machines: Measuring Information Processing in Natural Systems

Karl Young
Stanford Linear Accelerator
Stanford University
kyoung@slac.stanford.edu

This talk will focus on an attempt to define an intrinsic measure of information processing in physical systems (also known as complexity in some circles) and discuss ways of measuring it. Following some brief speculation about the types of system to which this might be fruitfully applied, an example will be presented. This example draws on methods from computation theory and grammar induction to classify simulated data from a simple nonlinear system.



Date: Thurs., Apr 12

Time: 4:15-5:30PM

Place: Cordura 100


Return to the seminar schedule