Faster solutions of the inverse pairwise Ising problem

Abstract

Recent work has shown that probabilistic models based on pairwise interactions-in the simplest case, the Ising model-provide surprisingly accurate descriptions of experiments on real biological networks ranging from neurons to genes. Finding these models requires us to solve an inverse problem: given experimentally measured expectation values, what are the parameters of the underlying Hamiltonian? This problem sits at the intersection of statistical physics and machine learning, and we suggest that more efficient solutions are possible by merging ideas from the two fields. We use a combination of recent coordinate descent algorithms with an adaptation of the histogram Monte Carlo method, and implement these techniques to take advantage of the sparseness found in data on real neurons. The resulting algorithm learns the parameters of an Ising model describing a network of forty neurons within a few minutes. This opens the possibility of analyzing much larger data sets now emerging, and thus testing hypotheses about the collective behaviors of these networks.

References in corpus (3)

• Weak pairwise correlations imply strongly correlated network states in a neural population
• Network information and connected correlations
• Ising models for networks of real neurons

Cited by in corpus (26)

• Are biological systems poised at criticality?
• Searching for collective behavior in a network of real neurons
• Statistical physics of pairwise probability models
• Stimulus-dependent maximum entropy models of neural population codes
• Spin glass models for a network of real neurons
• Toward a statistical mechanics of four letter words
• Signatures of criticality arise in simple neural population models with correlations
• Searching for simplicity: Approaches to the analysis of neurons and behavior
• Parameters estimation for spatio-temporal maximum entropy distributions: application to neural spike trains
• Learning Maximum Entropy Models from finite size datasets: a fast Data-Driven algorithm allows sampling from the posterior distribution
• A tractable method for describing complex couplings between neurons and population rate
• A Sampling-Based Approach to Computing Equilibria in Succinct Extensive-Form Games
• Ensemble Inhibition and Excitation in the Human Cortex: an Ising Model Analysis with Uncertainties
• Pairwise maximum-entropy models and their Glauber dynamics: bimodality, bistability, non-ergodicity problems, and their elimination via inhibition
• The Structured `Low Temperature' Phase of the Retinal Population Code
• Signatures of brain criticality unveiled by maximum entropy analysis across cortical states
• A statistical physics approach to learning curves for the Inverse Ising problem
• Minimum and maximum entropy distributions for binary systems with known means and pairwise correlations
• Maximum Entropy Principle Analysis in Network Systems with Short-time Recordings
• Time window to constrain the corner value of the global seismic-moment distribution
• Pairwise interactions origin of entropy functions
• Inverse problems in spin models
• Exponential Reduction in Sample Complexity with Learning of Ising Model Dynamics
• Convenient Interface to Inverse Ising (ConIII): A Python 3 Package for Solving Ising-Type Maximum Entropy Models
• GNisi: A graph network for reconstructing Ising models from multivariate binarized data
• Learning Continuous Occupancy Maps with the Ising Process Model