Summary
International Symposium on Nonlinear Theory and its Applications
2005
Session Number:2-1-3
Session:
Number:2-1-3-4
Learning Methods for Dynamic Neural Networks
Emmanuel Dauce, Hedi Soula, Guillaume Beslon,
pp.598-601
Publication Date:2005/10/18
Online ISSN:2188-5079
Summary:
In the framework of dynamic neural networks, learning refers to the slow process by which a neural network modi?es its own structure under the in?uence of environmental pressure. Our simulations take place on large random recurrent neural networks (RRNNs). We present several results obtained with the use of a TD (temporal di?erence) and STDP (Spike-Time Dependent Plasticity) rule. First, we show that under some conditions, those learning rules give rise to an increase of the neurons synchronization, which can be interpreted as the crossing of a bifurcation line between non-synchronized and synchronized regimes. Second, we present various results obtained in control, under a reinforcement learning paradigm: inverted pendulum control and obstacle avoidance.