Ax: A Platform for Adaptive Experimentation

Download PDF

Miles Olson, Elizabeth Santorella, Louis C. Tiao, Sait Cakmak, David Eriksson, Mia Garrard, Sam Daulton, Maximilian Balandat, Eytan Bakshy, Elena Kashtelyan, Zhiyuan Jerry Lin, Sebastian Ament, Bernard Beckerman, Eric Onofrey, Paschal Igusti, Cristian Lara, Benjamin Letham, Cesar Cardoso, Shiyun Sunny Shen, Andy Chenyuan Lin et al. (1 additional authors not shown)

Published: 03 Jun 2025, Last Modified: 03 Jun 2025AutoML 2025 ABCD TrackEveryoneRevisionsBibTeXCC BY 4.0
Confirmation: our paper adheres to reproducibility best practices. In particular, we confirm that all important details required to reproduce results are described in the paper,, the authors agree to the paper being made available online through OpenReview under a CC-BY 4.0 license (https://creativecommons.org/licenses/by/4.0/), and, the authors have read and commit to adhering to the AutoML 2025 Code of Conduct (https://2025.automl.cc/code-of-conduct/).
Reproducibility: zip
TL;DR: Ax is a versatile, extensive, robust platform for adaptive experimentation, including AutoML
Abstract: Optimizing industry-scale machine learning systems involves solving several resource-intensive black-box optimization problems. Adaptive experimentation can substantially improve the sample-efficiency of such tasks compared with naïve baselines (like grid or random search) by utilizing surrogate models and sequential optimization algorithms. In this paper, we present Ax (https://ax.dev), an open-source platform for adaptive experimentation. Ax is highly extensible and robust, and has been used at scale for production tasks at Meta since 2019. We discuss Ax's use, design, and analyze its performance. We demonstrate that, off-of-the-shelf, Ax achieves state-of-the-art performance on a broad range of synthetic and real-world black-box optimization tasks across machine learning, engineering, and scientific applications.
Submission Number: 9