English

L$_1$ Regularization for Reconstruction of a non-equilibrium Ising Model

Methodology 2012-11-19 v1 Disordered Systems and Neural Networks Data Analysis, Statistics and Probability

Abstract

The couplings in a sparse asymmetric, asynchronous Ising network are reconstructed using an exact learning algorithm. L1_1 regularization is used to remove the spurious weak connections that would otherwise be found by simply minimizing the minus likelihood of a finite data set. In order to see how L1_1 regularization works in detail, we perform the calculation in several ways including (1) by iterative minimization of a cost function equal to minus the log likelihood of the data plus an L1_1 penalty term, and (2) an approximate scheme based on a quadratic expansion of the cost function around its minimum. In these schemes, we track how connections are pruned as the strength of the L1_1 penalty is increased from zero to large values. The performance of the methods for various coupling strengths is quantified using ROC curves.

Keywords

Cite

@article{arxiv.1211.3671,
  title  = {L$_1$ Regularization for Reconstruction of a non-equilibrium Ising Model},
  author = {Hong-Li Zeng and John Hertz and Yasser Roudi},
  journal= {arXiv preprint arXiv:1211.3671},
  year   = {2012}
}
R2 v1 2026-06-21T22:39:06.783Z