Diffusion methods for classification with pairwise relationships

DOI10.1090/QAM/1540MaRDI QIDQ5233998zbMATH OpenOpenAlexWikidataFDO

Authors Pedro F. Felzenszwalb, B. F. Svaiter

Publication date 9 September 2019

Published in Quarterly of Applied Mathematics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1505.06072

zbMATH Keywords

classification problems with pairwise relationshipsnumerical experiments

Mathematics Subject Classification ID

Random walks on graphs (05C81) Graph algorithms (graph-theoretic aspects) (05C85) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Computing methodologies for image processing (68U10)

Abstract: We define two algorithms for propagating information in classification problems with pairwise relationships. The algorithms are based on contraction maps and are related to non-linear diffusion and random walks on graphs. The approach is also related to message passing algorithms, including belief propagation and mean field methods. The algorithms we describe are guaranteed to converge on graphs with arbitrary topology. Moreover they always converge to a unique fixed point, independent of initialization. We prove that the fixed points of the algorithms under consideration define lower-bounds on the energy function and the max-marginals of a Markov random field. The theoretical results also illustrate a relationship between message passing algorithms and value iteration for an infinite horizon Markov decision process. We illustrate the practical application of the algorithms under study with numerical experiments in image restoration, stereo depth estimation and binary classification on a grid.

Recommendations

• Approximation algorithms for classification problems with pairwise relationships, metric labeling and Markov random fields
• Probabilistic Relaxation Labeling by Fokker-Planck Diffusion on a Graph
• Convex combination belief propagation
• MAP Estimation Via Agreement on Trees: Message-Passing and Linear Programming
• On the optimality of solutions of the max-product belief-propagation algorithm in arbitrary graphs

Cites work

• Approximation algorithms for classification problems with pairwise relationships, metric labeling and Markov random fields
• Distance transforms of sampled functions
• Dynamic programming and optimal control. Vol. 1.
• Graphical models, exponential families, and variational inference
• scientific article; zbMATH DE number 3986503 (Why is no real title available?)
• scientific article; zbMATH DE number 3513115 (Why is no real title available?)
• scientific article; zbMATH DE number 2062604 (Why is no real title available?)
• MAP Estimation Via Agreement on Trees: Message-Passing and Linear Programming
• Message-passing for graph-structured linear programs: proximal methods and rounding schemes
• Nonserial dynamic programming
• Probabilistic graphical models.
• Pseudo-Boolean optimization
• Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images

Cited in

(3)

• Convex combination belief propagation
• Diffuse Interface Models on Graphs for Classification of High Dimensional Data
• Consistent procedures for multiclass classification of discrete diffusion paths

This page was built for publication: Diffusion methods for classification with pairwise relationships

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5233998)