Title :
Hebbian Learning using Fixed Weight Evolved Dynamical `Neural´ Networks
Author :
Izquierdo-Torres, Eduardo ; Harvey, Inman
Author_Institution :
Centre for Computational Neurosci. & Robotics, Sussex Univ., Brighton
Abstract :
We evolve small continuous-time recurrent neural networks with fixed weights that perform Hebbian learning behavior. We describe the performance of the best and smallest successful system, providing an in-depth analysis of its evolved mechanisms. Learning is shown to arise from the interaction between the multiple timescale dynamics. In particular, we show how the fast-time dynamics alter the slow-time dynamics, which in turn shapes the local behavior around the equilibrium points of the fast components by acting as a parameter to them
Keywords :
Hebbian learning; recurrent neural nets; Hebbian learning; continuous-time recurrent neural networks; fast-time dynamics; fixed weight evolved dynamical neural networks; multiple timescale dynamics; slow-time dynamics; Circuits; Computer networks; Hebbian theory; Neurons; Nonlinear systems; Performance analysis; Psychology; Recurrent neural networks; Robots; Shape;
Conference_Titel :
Artificial Life, 2007. ALIFE '07. IEEE Symposium on
Conference_Location :
Honolulu, HI
Print_ISBN :
1-4244-0701-X
DOI :
10.1109/ALIFE.2007.367822