Toward Optimal Run Racing: Application to Deep Learning Calibration

Olivier Bousquet (1) , Sylvain Gelly (1) , Karol Kurach (1) , Marc Schoenauer (2, 3) , Michèle Sebag (3, 2) , Olivier Teytaud (1) , Damien Vincent (1)

1 Google Brain
2 TAU - TAckling the Underspecified
3 LRI - Laboratoire de Recherche en Informatique

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.

Domaines: Intelligence artificielle [cs.AI]

Olivier Bousquet, Sylvain Gelly, Karol Kurach, Marc Schoenauer, Michèle Sebag, et al.. Toward Optimal Run Racing: Application to Deep Learning Calibration. 2017. ⟨hal-01634381⟩