Sparse Hopfield network reconstruction with ${\ell }_1$ regularization

Abstract

We propose an efficient strategy to infer sparse Hopfield network based on magnetizations and pairwise correlations measured through Glauber samplings. This strategy incorporates the ${\ell }_1$ regularization into the Bethe approximation by a quadratic approximation to the log-likelihood, and is able to further reduce the inference error of the Bethe approximation without the regularization. The optimal regularization parameter is observed to be of the order of $M^{-\nu }$ where $M$ is the number of independent samples. The value of the scaling exponent depends on the performance measure. $\nu \simeq 0.5001$ for root mean squared error measure while $\nu \simeq 0.2743$ for misclassification rate measure. The efficiency of this strategy is demonstrated for the sparse Hopfield model, but the method is generally applicable to other diluted mean field models. In particular, it is simple in implementation without heavy computational cost.

9 pages, 3 figures, Eur. Phys. J. B (in press)

References in corpus (7)

• Direct-coupling analysis of residue co-evolution captures native contacts across many protein families
• High-dimensional Ising model selection using ${\ell }_1$-regularized logistic regression
• Small-correlation expansions for the inverse Ising problem
• Adaptive cluster expansion for the inverse Ising problem: convergence, algorithm and tests
• The Bethe approximation for solving the inverse Ising problem: a comparison with other inference methods
• Bethe-Peierls approximation and the inverse Ising model
• Inference algorithms for gene networks: a statistical mechanics analysis

Cited by in corpus (3)

• Causality inference in stochastic systems from neurons to currencies: Profiting from small sample size
• Clustering of neural codewords revealed by a first-order phase transition
• Theory of population coupling and applications to describe high order correlations in large populations of interacting neurons