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-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. |
Year | DOI | Venue |
---|---|---|
2010 | 10.1103/PhysRevE.83.036114 | PHYSICAL REVIEW E |
Keywords | Field | DocType |
neural network | Attractor,Topology,Assortative mixing,Assortativity,Nerve cells,Theoretical computer science,Robustness (computer science),Artificial neural network,Classical mechanics,Mathematics | Journal |
Volume | Issue | ISSN |
83 | 3 | 1539-3755 |
Citations | PageRank | References |
4 | 0.56 | 0 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Sebastiano de Franciscis | 1 | 8 | 2.32 |
Samuel Johnson | 2 | 4 | 1.23 |
Joaquín J. Torres | 3 | 142 | 19.57 |