English

Fast Exact Matrix Completion: A Unified Optimization Framework for Matrix Completion

Machine Learning 2021-01-05 v2 Optimization and Control Methodology Machine Learning

Abstract

We formulate the problem of matrix completion with and without side information as a non-convex optimization problem. We design fastImpute based on non-convex gradient descent and show it converges to a global minimum that is guaranteed to recover closely the underlying matrix while it scales to matrices of sizes beyond 105×10510^5 \times 10^5. We report experiments on both synthetic and real-world datasets that show fastImpute is competitive in both the accuracy of the matrix recovered and the time needed across all cases. Furthermore, when a high number of entries are missing, fastImpute is over 75%75\% lower in MAPE and 1515 times faster than current state-of-the-art matrix completion methods in both the case with side information and without.

Keywords

Cite

@article{arxiv.1910.09092,
  title  = {Fast Exact Matrix Completion: A Unified Optimization Framework for Matrix Completion},
  author = {Dimitris Bertsimas and Michael Lingzhi Li},
  journal= {arXiv preprint arXiv:1910.09092},
  year   = {2021}
}
R2 v1 2026-06-23T11:49:16.148Z