Javascript must be enabled
Jake Bouvrie : Learning and Synchronization in Stochastic Neural Ensembles
We consider a biological learning model composed of coupled stochastic neural ensembles obeying a nonlinear gradient dynamics. The dynamics optimize a simple error criterion involving noisy observations provided by the environment, leading to a function that can be used to make decisions in the future. The uncertainty of the resulting decision function is characterized, and shown to be controlled in large part by trading off coupling strength (and/or network topology) against the ambient neuronal noise. Further connections with classical regularization notions in statistical learning theory will also be explored.
- Category: Graduate/Faculty Seminar
- Duration: 01:34:44
- Date: April 23, 2010 at 4:25 PM
- Views: 107
- Tags: seminar, Graduate/faculty Seminar
0 Comments