SubStrat: A Subset-Based Strategy for Faster AutoML
Teddy Lazebnik, Amit Somech, Abraham Itzhak Weinberg
TL;DR
SubStrat tackles AutoML's high computational cost on large datasets by introducing a subset-based strategy that preserves data characteristics using a dataset-measure, implemented as a measure-preserving data-subset (DST) discovered with a genetic algorithm. It first runs the target AutoML tool on the DST to obtain an intermediate configuration $M'$ and then fine-tunes the result by a constrained AutoML pass on the full dataset to produce $M_{sub}$. Experiments with Auto-Sklearn and TPOT across 10 datasets show substantial time reductions (about $79\%$ on average) with minimal accuracy loss (relative accuracy well above $95\%$ on most tasks). Compared to baselines such as random DSTs and standard row/column sampling, SubStrat consistently achieves higher speedups while preserving predictive performance, highlighting the practical value of data-size reduction in AutoML.
Abstract
Automated machine learning (AutoML) frameworks have become important tools in the data scientists' arsenal, as they dramatically reduce the manual work devoted to the construction of ML pipelines. Such frameworks intelligently search among millions of possible ML pipelines - typically containing feature engineering, model selection and hyper parameters tuning steps - and finally output an optimal pipeline in terms of predictive accuracy. However, when the dataset is large, each individual configuration takes longer to execute, therefore the overall AutoML running times become increasingly high. To this end, we present SubStrat, an AutoML optimization strategy that tackles the data size, rather than configuration space. It wraps existing AutoML tools, and instead of executing them directly on the entire dataset, SubStrat uses a genetic-based algorithm to find a small yet representative data subset which preserves a particular characteristic of the full data. It then employs the AutoML tool on the small subset, and finally, it refines the resulted pipeline by executing a restricted, much shorter, AutoML process on the large dataset. Our experimental results, performed on two popular AutoML frameworks, Auto-Sklearn and TPOT, show that SubStrat reduces their running times by 79% (on average), with less than 2% average loss in the accuracy of the resulted ML pipeline.