Mean-field approximation, convex hierarchies, and the optimality of correlation rounding: a unified perspective

  title={Mean-field approximation, convex hierarchies, and the optimality of correlation rounding: a unified perspective},
  author={Vishesh Jain and Frederic Koehler and Andrej Risteski},
  journal={Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing},
The free energy is a key quantity of interest in Ising models, but unfortunately, computing it in general is computationally intractable. Two popular (variational) approximation schemes for estimating the free energy of general Ising models (in particular, even in regimes where correlation decay does not hold) are: (i) the mean-field approximation with roots in statistical physics, which estimates the free energy from below, and (ii) hierarchies of convex relaxations with roots in theoretical… 

Mean field approximations via log-concavity

A new approach to deriving quantitative mean field approximations for any probability measure P on R with density proportional to e, for f strongly concave is proposed, based primarily on functional inequalities and the notion of displacement convexity from optimal transport.

Sampling Approximately Low-Rank Ising Models: MCMC meets Variational Methods

We consider Ising models on the hypercube with a general interaction matrix J , and give a polynomial time sampling algorithm when all but O(1) eigenvalues of J lie in an interval of length one, a

A transportation approach to the mean-field approximation

  • F. Augeri
  • Mathematics, Computer Science
    Probability Theory and Related Fields
  • 2021
These transportation-entropy inequalities which are saturated by measures such that their log-density with respect to the background measure is an affine function, in the setting of the uniform measure on the discrete hypercube and the exponential measure imply a strong integrability inequality for Bernoulli and exponential processes.

Improved Product-state Approximation Algorithms for Quantum Local Hamiltonians

New techniques to find classical, additive error productstate approximations for these quantities on certain families of Quantum k-Local Hamiltonians, including those which are either dense, have low threshold rank, or are defined on a sparse graph that excludes a fixed minor are developed.

Lower tails via relative entropy

We show that the naive mean-field approximation correctly predicts the leading term of the logarithmic lower tail probabilities for the number of copies of a given subgraph in G ( n,p ) and of

Fast Convergence of Belief Propagation to Global Optima: Beyond Correlation Decay

BP converges quickly to the global optimum of the Bethe free energy for Ising models on arbitrary graphs, as long as the Ising model is \emph{ferromagnetic} (i.e. neighbors prefer to be aligned).

PR ] 3 1 M ay 2 01 9 Taming correlations through entropy-efficient measure decompositions with applications to mean-field approximation

A general bound is derived for the mean-field approximation of Ising and Potts models, which is in a sense dimension free, in both continuous and discrete settings.

Nonlinear large deviation bounds with applications to traces of Wigner matrices and cycles counts in Erd\"os-Renyi graphs

It is proved that general nonlinear large deviation estimates similar to Chatterjee-Dembo's original bounds except that they do not require any second order smoothness are valid for a broad class of distributions.

Nonlinear large deviation bounds with applications to Wigner matrices and sparse Erdős–Rényi graphs

This work proves general nonlinear large deviation estimates similar to Chatterjee-Dembo’s original bounds except that they do not require any second order smoothness and are valid for a broad class of distributions.

Variational Inference in high-dimensional linear regression

This work derives a limiting infinite dimensional variational formula for the log normalizing constant of the posterior distribution of high-dimensional Bayesian linear regression with product priors and establishes that under an additional “separation" condition, the variational problem has a unique optimizer, and this optimizer governs the probabilistic properties ofThe posterior distribution.



How to calculate partition functions using convex programming hierarchies: provable bounds for variational methods

This work considers dense and low threshold rank graphs, and interestingly, the reason the approach works on these types of graphs is because local correlations propagate to global correlations – completely the opposite of algorithms based on correlation decay.

The Computational Hardness of Counting in Two-Spin Models on d-Regular Graphs

  • A. SlyNike Sun
  • Mathematics
    2012 IEEE 53rd Annual Symposium on Foundations of Computer Science
  • 2012
The proof establishes that the normalized log-partition function of any two-spin system on bipartite locally tree-like graphs converges to a limiting ``free energy density'' which coincides with the (non-rigorous) Be the prediction of statistical physics.

Graphical Models, Exponential Families, and Variational Inference

The variational approach provides a complementary alternative to Markov chain Monte Carlo as a general source of approximation methods for inference in large-scale statistical models.

Quick Approximation to Matrices and Applications

The matrix approximation is generalized to multi-dimensional arrays and from that derive approximation algorithms for all dense Max-SNP problems and the Regularity Lemma is derived.

Decomposition of mean-field Gibbs distributions into product measures

We show that under a low complexity condition on the gradient of a Hamiltonian, Gibbs distributions on the Boolean hypercube are approximate mixtures of product measures whose probability vectors are

Universality of the mean-field for the Potts model

We consider the Potts model with q colors on a sequence of weighted graphs with adjacency matrices $$A_n$$An, allowing for both positive and negative weights. Under a mild regularity condition on

Bethe States of Random Factor Graphs

It is verified that the Gibbs measure can be decomposed into a moderate number of Bethe states, subsets of the state space in which both short and long range correlations of the measure take a simple form.

Lasserre Hierarchy, Higher Eigenvalues, and Approximation Schemes for Graph Partitioning and Quadratic Integer Programming with PSD Objectives

  • V. GuruswamiA. Sinop
  • Computer Science, Mathematics
    2011 IEEE 52nd Annual Symposium on Foundations of Computer Science
  • 2011
An approximation scheme for optimizing certain Quadratic Integer Programming problems with positive semi definite objective functions and global linear constraints is presented, and an algorithm for independent sets in graphs that performs well when the Laplacian does not have too many eigenvalues bigger than $1+o(1).

Ising models on locally tree-like graphs

We consider Ising models on graphs that converge locally to trees. Examples include random regular graphs with bounded degree and uniformly random graphs with bounded average degree. We prove that

The statistics of Curie-Weiss models

LetSn denote the random total magnetization of ann-site Curie-Weiss model, a collection ofn (spin) random variables with an equal interaction of strength 1/n between each pair of spins. The