English

Probabilistic Rollouts for Learning Curve Extrapolation Across Hyperparameter Settings

Machine Learning 2019-10-11 v1 Machine Learning

Abstract

We propose probabilistic models that can extrapolate learning curves of iterative machine learning algorithms, such as stochastic gradient descent for training deep networks, based on training data with variable-length learning curves. We study instantiations of this framework based on random forests and Bayesian recurrent neural networks. Our experiments show that these models yield better predictions than state-of-the-art models from the hyperparameter optimization literature when extrapolating the performance of neural networks trained with different hyperparameter settings.

Keywords

Cite

@article{arxiv.1910.04522,
  title  = {Probabilistic Rollouts for Learning Curve Extrapolation Across Hyperparameter Settings},
  author = {Matilde Gargiani and Aaron Klein and Stefan Falkner and Frank Hutter},
  journal= {arXiv preprint arXiv:1910.04522},
  year   = {2019}
}
R2 v1 2026-06-23T11:39:41.889Z