ICML 2019oral27 citations

CapsAndRuns: An Improved Method for Approximately Optimal Algorithm Configuration

Gellert Weisz, Andras Gyorgy, Csaba Szepesvari

Abstract

We consider the problem of configuring general-purpose solvers to run efficiently on problem instances drawn from an unknown distribution, a problem of major interest in solver autoconfiguration. Following previous work, we focus on designing algorithms that find a configuration with near-optimal expected capped runtime while doing the least amount of work, with the cap chosen in a configuration-specific way so that most instances are solved. In this paper we present a new algorithm, CapsAndRuns, which finds a near-optimal configuration while using time that scales (in a problem dependent way) with the optimal expected capped runtime, significantly strengthening previous results which could only guarantee a bound that scaled with the potentially much larger optimal expected uncapped runtime. The new algorithm is simpler and more intuitive than the previous methods: first it estimates the optimal runtime cap for each configuration, then it uses a Bernstein race to find a near optimal configuration given the caps. Experiments verify that our method can significantly outperform its competitors.

BibTeX
@InProceedings{pmlr-v97-weisz19a,
  title = 	 {{C}aps{A}nd{R}uns: An Improved Method for Approximately Optimal Algorithm Configuration},
  author =       {Weisz, Gellert and Gyorgy, Andras and Szepesvari, Csaba},
  booktitle = 	 {Proceedings of the 36th International Conference on Machine Learning},
  pages = 	 {6707--6715},
  year = 	 {2019},
  editor = 	 {Chaudhuri, Kamalika and Salakhutdinov, Ruslan},
  volume = 	 {97},
  series = 	 {Proceedings of Machine Learning Research},
  month = 	 {09--15 Jun},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v97/weisz19a/weisz19a.pdf},
  url = 	 {https://proceedings.mlr.press/v97/weisz19a.html},
  abstract = 	 {We consider the problem of configuring general-purpose solvers to run efficiently on problem instances drawn from an unknown distribution, a problem of major interest in solver autoconfiguration. Following previous work, we focus on designing algorithms that find a configuration with near-optimal expected capped runtime while doing the least amount of work, with the cap chosen in a configuration-specific way so that most instances are solved. In this paper we present a new algorithm, CapsAndRuns, which finds a near-optimal configuration while using time that scales (in a problem dependent way) with the optimal expected capped runtime, significantly strengthening previous results which could only guarantee a bound that scaled with the potentially much larger optimal expected uncapped runtime. The new algorithm is simpler and more intuitive than the previous methods: first it estimates the optimal runtime cap for each configuration, then it uses a Bernstein race to find a near optimal configuration given the caps. Experiments verify that our method can significantly outperform its competitors.}
}
CapsAndRuns: An Improved Method for Approximately Optimal Algorithm Configuration · ICML 2019