English

Enhancing neural-network performance via assortativity

Disordered Systems and Neural Networks 2015-05-20 v1 Performance Biological Physics Neurons and Cognition

Abstract

The performance of attractor neural networks has been shown to depend crucially on the heterogeneity of the underlying topology. We take this analysis a step further by examining the effect of degree-degree correlations -- or assortativity -- on neural-network behavior. We make use of a method recently put forward for studying correlated networks and dynamics thereon, both analytically and computationally, which is independent of how the topology may have evolved. We show how the robustness to noise is greatly enhanced in assortative (positively correlated) neural networks, especially if it is the hub neurons that store the information.

Keywords

Cite

@article{arxiv.1012.1813,
  title  = {Enhancing neural-network performance via assortativity},
  author = {Sebastiano de Franciscis and Samuel Johnson and Joaquín J. Torres},
  journal= {arXiv preprint arXiv:1012.1813},
  year   = {2015}
}

Comments

9 pages, 7 figures

R2 v1 2026-06-21T16:55:31.409Z