Don't Waste Your Time: Early Stopping Cross-Validation

AutoML 2024 Conference Submission25 Authors

Published: 30 Apr 2024, Last Modified: 18 Jul 2024AutoML 2024EveryoneRevisionsBibTeXCC BY 4.0
Keywords: AutoML, Model Selection, HPO, Cross-Validation, Early Stopping, Bayesian Optimization, Overfitting
TL;DR: More effective model selection through early stopping cross-validation.
Abstract: State-of-the-art automated machine learning systems for tabular data often employ cross-validation; ensuring that measured performances generalize to unseen data, or that subsequent ensembling does not overfit. However, using k-fold cross-validation instead of holdout validation drastically increases the computational cost of validating a single configuration. While ensuring better generalization and, by extension, better performance, the additional cost is often prohibitive for effective model selection within a time budget. We aim to make model selection with cross-validation more effective. Therefore, we study early stopping the process of cross-validation during model selection. We investigate the impact of early stopping on random search for two algorithms, MLP and random forest, across 36 classification datasets. We further analyze the impact of the number of folds by considering 3-, 5-, and 10-folds. In addition, we ablate the impact of early stopping on Bayesian optimization and also repeated cross-validation. Our exploratory study shows that even a simple-to-understand and easy-to-implement method consistently allows model selection to converge faster; in ~94% of all datasets, on average by 214%. Moreover, stopping cross-validation enables model selection to explore the search space more exhaustively by considering +167% configurations on average, while also obtaining better overall performance.
Submission Checklist: Yes
Broader Impact Statement: Yes
Paper Availability And License: Yes
Code Of Conduct: Yes
Optional Meta-Data For Green-AutoML: This blue field is just for structuring purposes and cannot be filled.
CPU Hours: 53856
GPU Hours: 0
TPU Hours: 0
Evaluation Metrics: No
Community Implementations:
Submission Number: 25