Collaborative Filtering in a Non-Uniform World: Learning with the Weighted Trace Norm

TL;DR

This paper introduces a weighted trace-norm regularizer for matrix completion that performs better under non-uniform sampling, demonstrated by improved results on the Netflix dataset.

Contribution

The paper proposes a novel weighted trace-norm regularizer that effectively handles non-uniform sampling in matrix completion tasks.

Findings

Weighted trace-norm improves matrix completion accuracy.

Significant gains on Netflix dataset with non-uniform sampling.

Regularizer outperforms standard trace-norm in experiments.

Abstract

We show that matrix completion with trace-norm regularization can be significantly hurt when entries of the matrix are sampled non-uniformly. We introduce a weighted version of the trace-norm regularizer that works well also with non-uniform sampling. Our experimental results demonstrate that the weighted trace-norm regularization indeed yields significant gains on the (highly non-uniformly sampled) Netflix dataset.