Large pseudocounts and L[subscript 2]-norm penalties are necessary for the mean-field inference of Ising and Potts models
Name
Barton-2014-Large pseudocounts.pdf
Size
2.37 MB
Format
Adobe PDF
Checksum (MD5)
961a05e26e407ddd14e1df705ee9fc4c
Author(s)
Cocco, S.
De Leonardis, E.
Monasson, R.
Barton, John P.
Date Issued
July 2014
Journal
Physical Review E
Publisher
American Physical Society
Citation
Barton, J. P., S. Cocco, E. De Leonardis, and R. Monasson. “Large Pseudocounts and L[subscript 2]-Norm Penalties Are Necessary for the Mean-Field Inference of Ising and Potts Models.” Phys. Rev. E 90, no. 1 (July 2014). © 2014 American Physical Society
Version
Final published version
Abstract
The mean-field (MF) approximation offers a simple, fast way to infer direct interactions between elements in a network of correlated variables, a common, computationally challenging problem with practical applications in fields ranging from physics and biology to the social sciences. However, MF methods achieve their best performance with strong regularization, well beyond Bayesian expectations, an empirical fact that is poorly understood. In this work, we study the influence of pseudocount and L[subscript 2]-norm regularization schemes on the quality of inferred Ising or Potts interaction networks from correlation data within the MF approximation. We argue, based on the analysis of small systems, that the optimal value of the regularization strength remains finite even if the sampling noise tends to zero, in order to correct for systematic biases introduced by the MF approximation. Our claim is corroborated by extensive numerical studies of diverse model systems and by the analytical study of the m-component spin model for large but finite m. Additionally, we find that pseudocount regularization is robust against sampling noise and often outperforms L[subscript 2]-norm regularization, particularly when the underlying network of interactions is strongly heterogeneous. Much better performances are generally obtained for the Ising model than for the Potts model, for which only couplings incoming onto medium-frequency symbols are reliably inferred.
MIT Department
Massachusetts Institute of Technology. Department of Chemical Engineering
Ragon Institute of MGH, MIT and Harvard
Terms of Use
Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.
Persistent DSpace Link
DOI of Published Version
https://doi.org/10.1103/PhysRevE.90.012132
