Uncertainty quantification for Markov random fields

DOI10.1137/20M1374614zbMATH Open1473.62197arXiv2009.00038OpenAlexW3208339941MaRDI QIDQ5158928FDOQ5158928

Authors: P. Birmpa, Markos Katsoulakis 

Publication date: 26 October 2021

Published in: SIAM/ASA Journal on Uncertainty Quantification (Search for Journal in Brave)

Abstract: We present an information-based uncertainty quantification method for general Markov Random Fields. Markov Random Fields (MRF) are structured, probabilistic graphical models over undirected graphs, and provide a fundamental unifying modeling tool for statistical mechanics, probabilistic machine learning, and artificial intelligence. Typically MRFs are complex and high-dimensional with nodes and edges (connections) built in a modular fashion from simpler, low-dimensional probabilistic models and their local connections; in turn, this modularity allows to incorporate available data to MRFs and efficiently simulate them by leveraging their graph-theoretic structure. Learning graphical models from data and/or constructing them from physical modeling and constraints necessarily involves uncertainties inherited from data, modeling choices, or numerical approximations. These uncertainties in the MRF can be manifested either in the graph structure or the probability distribution functions, and necessarily will propagate in predictions for quantities of interest. Here we quantify such uncertainties using tight, information based bounds on the predictions of quantities of interest; these bounds take advantage of the graphical structure of MRFs and are capable of handling the inherent high-dimensionality of such graphical models. We demonstrate our methods in MRFs for medical diagnostics and statistical mechanics models. In the latter, we develop uncertainty quantification bounds for finite size effects and phase diagrams, which constitute two of the typical predictions goals of statistical mechanics modeling.

Full work available at URL: https://arxiv.org/abs/2009.00038

Recommendations

• scientific article; zbMATH DE number 1247560
• Mixed Markov models
• Markov random fields and their applications
• Prior specification of neighbourhood and interaction structure in binary Markov random fields
• Model Uncertainty and Correctability for Directed Graphical Models

zbMATH Keywords

uncertainty quantificationinformation theoryMarkov random fieldslong range interactionsprobabilistic inequalities

Mathematics Subject Classification ID

Probabilistic graphical models (62H22) Measures of information, entropy (94A17) Lattice systems (Ising, dimer, Potts, etc.) and systems on graphs arising in equilibrium statistical mechanics (82B20)

Cites Work

• The elements of statistical learning. Data mining, inference, and prediction
• Graphical models, exponential families, and variational inference
• Statistical mechanics: theory and molecular simulation.
• Machine learning. A probabilistic perspective
• Title not available (Why is that?)
• Introduction to nonparametric estimation
• Probabilistic graphical models.
• Deep learning
• Modeling and Reasoning with Bayesian Networks
• Title not available (Why is that?)
• Probability on graphs. Random processes on graphs and lattices.
• Title not available (Why is that?)
• Existence of a phase-transition in a one-dimensional Ising ferromagnet
• A review of linear response theory for general differentiable dynamical systems
• Global sensitivity analysis: The primer
• The phase transition in the one-dimensional Ising model with \(1/r^2\) interaction energy
• Computation of failure probability subject to epistemic uncertainty
• Discontinuity of the magnetization in one-dimensional \(1/| x-y| ^ 2\) Ising and Potts models.
• Title not available (Why is that?)
• Scaling limits in statistical mechanics and microstructures in continuum mechanics
• Worst-Case Value-At-Risk and Robust Portfolio Optimization: A Conic Programming Approach
• Title not available (Why is that?)
• From equilibrium to nonequilibrium statistical mechanics. phase transitions and the Fourier law
• Noise propagation in hybrid models of nonlinear systems: the Ginzburg-Landau equation
• Markov logic. An interface layer for artificial intelligence. With contributions by Jesse Davis, Tuyen Huynh, Stanley Kok, Lilyana Mihalkova, Raymond J. Mooney, Aniruddh Nath, Hoifung Poon, Matthew Richardson, Parag Singla, Marc Sumner, and Jue Wang
• The low-temperature phase of Kac-Ising models
• Title not available (Why is that?)
• Statistical mechanics of disordered systems. A mathematical perspective.
• Statistical mechanics of lattice systems. A concrete mathematical introduction
• Study of a long range perturbation of a one-dimensional Kac model
• Robust risk measurement and model risk
• Path-space information bounds for uncertainty quantification and sensitivity analysis of stochastic dynamics
• Robust bounds on risk-sensitive functionals via Rényi divergence
• Scalable information inequalities for uncertainty quantification
• Distinguishing and integrating aleatoric and epistemic variation in uncertainty quantification
• Uncertainty quantification for Markov processes via variational principles and functional inequalities
• How Biased Is Your Model? Concentration Inequalities, Information and Model Bias
• GINNs: graph-informed neural networks for multiscale physics
• Causality and Bayesian network PDEs for multiscale representations of porous media
• Mutual information for explainable deep learning of multiscale systems
• Sensitivity analysis for rare events based on Rényi divergence
• Title not available (Why is that?)

Cited In (3)

• Title not available (Why is that?)
• Quantifying and Managing Uncertainty in Piecewise-Deterministic Markov Processes
• Model Uncertainty and Correctability for Directed Graphical Models

Uses Software

• DAKOTA
• ElemStatLearn
• PMTK

This page was built for publication: Uncertainty quantification for Markov random fields

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5158928)