Stability and Generalization for Randomized Coordinate Descent

Wang, Puyu; Wu, Liang; Lei, Yunwen

doi:10.24963/ijcai.2021/427

Stability and Generalization for Randomized Coordinate Descent

Puyu Wang, Liang Wu, Yunwen Lei

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 3104-3110.

https://doi.org/10.24963/ijcai.2021/427

Keywords

Machine Learning: Online Learning
Machine Learning: Learning Theory

Abstract

Randomized coordinate descent (RCD) is a popular optimization algorithm with wide applications in various machine learning problems, which motivates a lot of theoretical analysis on its convergence behavior. As a comparison, there is no work studying how the models trained by RCD would generalize to test examples. In this paper, we initialize the generalization analysis of RCD by leveraging the powerful tool of algorithmic stability. We establish argument stability bounds of RCD for both convex and strongly convex objectives, from which we develop optimal generalization bounds by showing how to early-stop the algorithm to tradeoff the estimation and optimization. Our analysis shows that RCD enjoys better stability as compared to stochastic gradient descent.

Puyu Wang, Liang Wu, and Yunwen Lei. 2021. Stability and Generalization for Randomized Coordinate Descent. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 3104–3110. https://doi.org/10.24963/ijcai.2021/427

@inproceedings{ijcai2021p427,
  title     = {Stability and Generalization for Randomized Coordinate Descent},
  author    = {Wang, Puyu and Wu, Liang and Lei, Yunwen},
  booktitle = {Proceedings of the Thirtieth International Joint Conference on
               Artificial Intelligence, {IJCAI-21}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Prof. Zhi-Hua Zhou},
  pages     = {3104--3110},
  year      = {2021},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2021/427},
  url       = {https://doi.org/10.24963/ijcai.2021/427},
}

TY  - CPAPER
TI  - Stability and Generalization for Randomized Coordinate Descent
AU  - Wang, Puyu
AU  - Wu, Liang
AU  - Lei, Yunwen
T2  - Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2021
SP  - 3104
EP  - 3110
DO  - 10.24963/ijcai.2021/427
UR  - https://doi.org/10.24963/ijcai.2021/427
ER  -