Adversarial Robust Low Rank Matrix Estimation: Compressed Sensing and Matrix Completion

TL;DR

This paper introduces a robust method for low rank matrix estimation under adversarial contamination, providing sharp error bounds for compressed sensing and matrix completion using a unified approach with Huber loss and nuclear norm.

Contribution

It proposes a novel unified approach combining Huber loss and nuclear norm for robust low rank matrix estimation, improving error bounds over previous methods.

Findings

Sharp estimation error bounds for matrix compressed sensing.

Effective handling of adversarial contamination in matrix completion.

Unified approach applicable to multiple low rank matrix estimation problems.

Abstract

We consider robust low rank matrix estimation as a trace regression when outputs are contaminated by adversaries. The adversaries are allowed to add arbitrary values to arbitrary outputs. Such values can depend on any samples. We deal with matrix compressed sensing, including lasso as a partial problem, and matrix completion, and then we obtain sharp estimation error bounds. To obtain the error bounds for different models such as matrix compressed sensing and matrix completion, we propose a simple unified approach based on a combination of the Huber loss function and the nuclear norm penalization, which is a different approach from the conventional ones. Some error bounds obtained in the present paper are sharper than the past ones.