PubMed Health⌕ Search

Biomedical subjects

O Kinouchi

Publications and source records attributed to O Kinouchi.

6 recordsLinked to original sources
PubMed2001

Deterministic walks in random media.

Deterministic walks over a random set of N points in one and two dimensions ( d = 1,2) are considered. Points ("cities") are randomly scattered in R(d) following a uniform distribution. A walker ("tourist"), at each time step, goes to the nearest neighbor city that has not been visited in the past tau steps. Each initial city leads to a different trajectory composed of a transient part and a final p-cycle attractor. Transient times (for d = 1,2) follow an exponential law with a tau-dependent decay time but the density of p cycles can be approximately described by D(p)proportional to p(-alpha(tau)). For tau>>1 and tau/N<<1, the exponent is independent of tau. Some analytical results are given for the d = 1 case.

Journal Article↗
PubMed2000

Optimal pruning in neural networks.

We study pruning strategies in simple perceptrons subjected to supervised learning. Our analytical results, obtained through the statistical mechanics approach to learning theory, are independent of the learning algorithm used in the training process. We calculate the post-training distribution P(J) of synaptic weights, which depends only on the overlap rho(0) achieved by the learning algorithm before pruning and the fraction kappa of relevant weights in the teacher network. From this distribution, we calculate the optimal pruning strategy for deleting small weights. The optimal pruning threshold grows from zero as straight theta(opt)(rho(0), kappa) approximately [rho(0)-rho(c)(kappa)](1/2) above some critical value rho(c)(kappa). Thus, the elimination of weak synapses enhances the network performance only after a critical learning period. Possible implications for biological pruning phenomena are discussed.

Algorithms↗
PubMed1999

Robustness of scale invariance in models with self-organized criticality.

A random-neighbor extremal stick-slip model is introduced. In the thermodynamic limit, the distribution of states has a simple analytical form and the mean avalanche size, as a function of the coupling parameter, is exactly calculable. The system is critical only at a special point J(c) in coupling parameter space. However, the critical region around this point, where approximate scale invariance holds, is very large, suggesting a mechanism for explaining the ubiquity of power laws in nature.

Journal Article↗