English

Toward Optimal Run Racing: Application to Deep Learning Calibration

Machine Learning 2017-06-21 v2

Abstract

This paper aims at one-shot learning of deep neural nets, where a highly parallel setting is considered to address the algorithm calibration problem - selecting the best neural architecture and learning hyper-parameter values depending on the dataset at hand. The notoriously expensive calibration problem is optimally reduced by detecting and early stopping non-optimal runs. The theoretical contribution regards the optimality guarantees within the multiple hypothesis testing framework. Experimentations on the Cifar10, PTB and Wiki benchmarks demonstrate the relevance of the approach with a principled and consistent improvement on the state of the art with no extra hyper-parameter.

Keywords

Cite

@article{arxiv.1706.03199,
  title  = {Toward Optimal Run Racing: Application to Deep Learning Calibration},
  author = {Olivier Bousquet and Sylvain Gelly and Karol Kurach and Marc Schoenauer and Michele Sebag and Olivier Teytaud and Damien Vincent},
  journal= {arXiv preprint arXiv:1706.03199},
  year   = {2017}
}
R2 v1 2026-06-22T20:14:51.036Z