Put CASH on Bandits: A Max K-Armed Problem for Automated Machine Learning

Amir Rezaei Balef, Claire Vernade, Katharina Eggensperger

Keywords: Multi Armed Bandit, Automated Machine Learning, Hyperparameter Optimization

TL;DR: A Max K-armed Bandit using assumptions derived from empirical data that handles short-tailed and skewed distributions to dynamically allocate resources to hyperparameter optimization runs.

Abstract: The Combined Algorithm Selection and Hyperparameter optimization (CASH) is a challenging resource allocation problem in the field of AutoML. We propose MaxUCB, a max $k$-armed bandit method to trade off exploring different model classes and conducting hyperparameter optimization. MaxUCB is specifically designed for the light-tailed and bounded reward distributions arising in this setting and, thus, provides an efficient alternative compared to classic max $k$-armed bandit methods assuming heavy-tailed reward distributions. We theoretically and empirically evaluate our method on four standard AutoML benchmarks, demonstrating superior performance over prior approaches. We make our code and data available at https://github.com/amirbalef/CASH_with_Bandits

Primary Area: General machine learning (supervised, unsupervised, online, active, etc.)

Submission Number: 10484