Stochastic Evolution of Neural Networks

F.I. Karpelevich, V.A. Malyshev, A.N. Rybko

1995, v.1, №1, 141-161


We study the dynamics of neural networks introduced by M. Cottrell for the case of symmetric connections ($a_{ij}=a_{ji}$). We study the structure of images that can be remembered by these networks and prove convergence to these images starting from approximate images.

Keywords: neural networks,random walks in the orthant,exit boundaries,Euler limit,processes with local interaction


Please log in or register to leave a comment

There are no comments yet