English

DID: Distributed Incremental Block Coordinate Descent for Nonnegative Matrix Factorization

Machine Learning 2018-02-27 v1 Artificial Intelligence Optimization and Control Machine Learning

Abstract

Nonnegative matrix factorization (NMF) has attracted much attention in the last decade as a dimension reduction method in many applications. Due to the explosion in the size of data, naturally the samples are collected and stored distributively in local computational nodes. Thus, there is a growing need to develop algorithms in a distributed memory architecture. We propose a novel distributed algorithm, called \textit{distributed incremental block coordinate descent} (DID), to solve the problem. By adapting the block coordinate descent framework, closed-form update rules are obtained in DID. Moreover, DID performs updates incrementally based on the most recently updated residual matrix. As a result, only one communication step per iteration is required. The correctness, efficiency, and scalability of the proposed algorithm are verified in a series of numerical experiments.

Keywords

Cite

@article{arxiv.1802.08938,
  title  = {DID: Distributed Incremental Block Coordinate Descent for Nonnegative Matrix Factorization},
  author = {Tianxiang Gao and Chris Chu},
  journal= {arXiv preprint arXiv:1802.08938},
  year   = {2018}
}

Comments

Accepted by AAAI 2018

R2 v1 2026-06-23T00:32:30.876Z