Tuning-Free Online Robust Principal Component Analysis through Implicit Regularization

TL;DR

This paper introduces a tuning-free online robust PCA method that leverages implicit regularization through modified gradient descent, eliminating the need for dataset-specific parameter tuning and improving scalability.

Contribution

It proposes a novel tuning-free OR-PCA approach using implicit regularization with modified gradient descent, enhancing scalability and performance.

Findings

Performs comparably or better than tuned OR-PCA on various datasets.

Eliminates the need for dataset-dependent parameter tuning.

Improves scalability for large datasets.

Abstract

The performance of the standard Online Robust Principal Component Analysis (OR-PCA) technique depends on the optimum tuning of the explicit regularizers and this tuning is dataset sensitive. We aim to remove the dependency on these tuning parameters by using implicit regularization. We propose to use the implicit regularization effect of various modified gradient descents to make OR-PCA tuning free. Our method incorporates three different versions of modified gradient descent that separately but naturally encourage sparsity and low-rank structures in the data. The proposed method performs comparable or better than the tuned OR-PCA for both simulated and real-world datasets. Tuning-free ORPCA makes it more scalable for large datasets since we do not require dataset-dependent parameter tuning.