Online Nonnegative Matrix Factorization With Outliers

DOI10.1109/TSP.2016.2620967zbMATH Open1416.65122arXiv1604.02634OpenAlexW2341265152MaRDI QIDQ4620564FDOQ4620564

Authors: Renbo Zhao, Vincent Y. F. Tan 

Publication date: 8 February 2019

Published in: IEEE Transactions on Signal Processing (Search for Journal in Brave)

Abstract: We propose a unified and systematic framework for performing online nonnegative matrix factorization in the presence of outliers. Our framework is particularly suited to large-scale data. We propose two solvers based on projected gradient descent and the alternating direction method of multipliers. We prove that the sequence of objective values converges almost surely by appealing to the quasi-martingale convergence theorem. We also show the sequence of learned dictionaries converges to the set of stationary points of the expected loss function almost surely. In addition, we extend our basic problem formulation to various settings with different constraints and regularizers. We also adapt the solvers and analyses to each setting. We perform extensive experiments on both synthetic and real datasets. These experiments demonstrate the computational efficiency and efficacy of our algorithms on tasks such as (parts-based) basis learning, image denoising, shadow removal and foreground-background separation.

Full work available at URL: https://arxiv.org/abs/1604.02634

Mathematics Subject Classification ID

Factorization of matrices (15A23) Positive matrices and their generalizations; cones of matrices (15B48)

Cited In (4)

• ORKM: online regularized \(K\)-means clustering for online multi-view data
• Title not available (Why is that?)
• Online Forecasting Matrix Factorization
• Momentum-based variance-reduced proximal stochastic gradient method for composite nonconvex stochastic optimization

This page was built for publication: Online Nonnegative Matrix Factorization With Outliers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4620564)