On the Convergence Rate of Decomposable Submodular Function Minimization

Abstract

Submodular functions describe a variety of discrete problems in machine learning, signal processing, and computer vision. However, minimizing submodular functions poses a number of algorithmic challenges. Recent work introduced an easy-to-use, parallelizable algorithm for minimizing submodular functions that decompose as the sum of "simple" submodular functions. Empirically, this algorithm performs extremely well, but no theoretical analysis was given. In this paper, we show that the algorithm converges linearly, and we provide upper and lower bounds on the rate of convergence. Our proof relies on the geometry of submodular polyhedra and draws on results from spectral graph theory.

17 pages, 3 figures

Cited by in corpus (7)

• Submodular Hypergraphs: p-Laplacians, Cheeger Inequalities and Spectral Clustering
• Computing exact minimum cuts without knowing the graph
• Structured Optimal Transport
• Convex Optimization for Parallel Energy Minimization
• Active-set Methods for Submodular Minimization Problems
• Approximate Decomposable Submodular Function Minimization for Cardinality-Based Components
• An Efficient Decomposition Framework for Discriminative Segmentation with Supermodular Losses