English

Subgradient Descent Learns Orthogonal Dictionaries

Machine Learning 2019-07-02 v2 Information Theory math.IT Optimization and Control Machine Learning

Abstract

This paper concerns dictionary learning, i.e., sparse coding, a fundamental representation learning problem. We show that a subgradient descent algorithm, with random initialization, can provably recover orthogonal dictionaries on a natural nonsmooth, nonconvex 1\ell_1 minimization formulation of the problem, under mild statistical assumptions on the data. This is in contrast to previous provable methods that require either expensive computation or delicate initialization schemes. Our analysis develops several tools for characterizing landscapes of nonsmooth functions, which might be of independent interest for provable training of deep networks with nonsmooth activations (e.g., ReLU), among numerous other applications. Preliminary experiments corroborate our analysis and show that our algorithm works well empirically in recovering orthogonal dictionaries.

Keywords

Cite

@article{arxiv.1810.10702,
  title  = {Subgradient Descent Learns Orthogonal Dictionaries},
  author = {Yu Bai and Qijia Jiang and Ju Sun},
  journal= {arXiv preprint arXiv:1810.10702},
  year   = {2019}
}
R2 v1 2026-06-23T04:52:06.602Z