Abstract:We study the problem of approximating the partition function of the ferromagnetic Ising model with both pairwise as well as higher order interactions (equivalently, in graphs as well as hypergraphs). Our approach is based on the classical Lee–Yang theory of phase transitions, along with a new Lee–Yang theorem for the Ising model with higher order interactions, and on an extension of ideas developed recently by Barvinok, and Patel and Regts that can be seen as an algorithmic realization of the Lee–Yang theory. Our first result is a deterministic polynomial time approximation scheme (an FPTAS) for the partition function in bounded degree graphs that is valid over the entire range of parameters β$$\beta $$ (the interaction) and λ$$\lambda $$ (the external field), except for the case λ=1$$\left| \lambda \right| =1$$ (the “zero-field” case). A polynomial time randomized approximation scheme (FPRAS) for all graphs and all β,λ$$\beta ,\lambda $$, based on Markov chain Monte Carlo simulation, has long been known. Unlike most other deterministic approximation algorithms for problems in statistical physics and counting, our algorithm does not rely on the “decay of correlations” property, but, as pointed out above, on Lee–Yang theory. This approach extends to the more general setting of the Ising model on hypergraphs of bounded degree and edge size, where no previous algorithms (even randomized) were known for a wide range of parameters. In order to achieve this latter extension, we establish a tight version of the Lee–Yang theorem for the Ising model on hypergraphs, improving a classical result of Suzuki and Fisher.

Intervention in undirected Ising graphs and the partition function

Efficient Exact Inference in Planar Ising Models

The Ising Partition Function: Zeros and Deterministic Approximation

Ising model partition-function computation as a weighted counting problem

Generalized Ising Model on a Scale-Free Network: An Interplay of Power Laws

Learning Planar Ising Models

Characterization and Greedy Learning of Interventional Markov Equivalence Classes of Directed Acyclic Graphs

Inverse Ising techniques to infer underlying mechanisms from data

Perturbation graphs, invariant prediction and causal relations in psychology

Distributional Invariances and Interventional Markov Equivalence for Mixed Graph Models

A Note on Ising Network Analysis with Missing Data

Interaction Screening: Efficient and Sample-Optimal Learning of Ising Models

The Erdős-Rényi Random Graph Conditioned on Every Component Being a Clique

Inference for a Large Directed Acyclic Graph with Unspecified Interventions

New algorithm for the Ising problem: partition function for finite lattice graphs

A Graph-Theoretic Approach to Randomization Tests of Causal Effects Under General Interference

Perturbation graphs, invariant causal prediction and causal relations in psychology

Ising's roots and the transfer-matrix eigenvalues

Ising model on a Galton-Watson tree with a sparse random external field