pith. sign in

arxiv: cond-mat/0512448 · v1 · submitted 2005-12-19 · ❄️ cond-mat.dis-nn · cond-mat.soft

Bump formation in a binary attractor neural network

classification ❄️ cond-mat.dis-nn cond-mat.soft
keywords formationnetworkwhenactivityanalyticalattractorbinarybump
0
0 comments X
read the original abstract

This paper investigates the conditions for the formation of local bumps in the activity of binary attractor neural networks with spatially dependent connectivity. We show that these formations are observed when asymmetry between the activity during the retrieval and learning is imposed. Analytical approximation for the order parameters is derived. The corresponding phase diagram shows a relatively large and stable region, where this effect is observed, although the critical storage and the information capacities drastically decrease inside that region. We demonstrate that the stability of the network, when starting from the bump formation, is larger than the stability when starting even from the whole pattern. Finally, we show a very good agreement between the analytical results and the simulations performed for different topologies of the network.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.