Topology and Computational Performance of Attractor Neural Networks
arXiv:cond-mat/0304021 · doi:10.1103/PhysRevE.68.047102
Abstract
To explore the relation between network structure and function, we studied the computational performance of Hopfield-type attractor neural nets with regular lattice, random, small-world and scale-free topologies. The random net is the most efficient for storage and retrieval of patterns by the entire network. However, in the scale-free case retrieval errors are not distributed uniformly: the portion of a pattern encoded by the subset of highly connected nodes is more robust and efficiently recognized than the rest of the pattern. The scale-free network thus achieves a very strong partial recognition. Implications for brain function and social dynamics are suggestive.
2 figures included. Submitted to Phys. Rev. Letters
References in corpus (5)
Cited by in corpus (12)
- Performance of networks of artificial neurons: The role of clustering
- Associative memory on a small-world neural network
- Analytic solution of attractor neural networks on scale-free graphs
- Modular organization enhances the robustness of attractor network dynamics
- Slow relaxation in the Ising model on a small-world network with strong long-range interactions
- Topology and Dynamics of Attractor Neural Networks: The Role of Loopiness
- Transient Dynamics of Sparsely Connected Hopfield Neural Networks with Arbitrary Degree Distributions
- Statistical-mechanical iterative algorithms on complex networks
- Maximum memory capacity on neural networks with short-term depression and facilitation
- Autoassociative Memory Retrieval and Spontaneous Activity Bumps in Small-World Networks of Integrate-and-Fire Neurons
- Transient dynamics for sequence processing neural networks: effect of degree distributions
- Optimising the topology of complex neural networks