English

Hyperparameter Optimization with Differentiable Metafeatures

Machine Learning 2021-02-09 v1

Abstract

Metafeatures, or dataset characteristics, have been shown to improve the performance of hyperparameter optimization (HPO). Conventionally, metafeatures are precomputed and used to measure the similarity between datasets, leading to a better initialization of HPO models. In this paper, we propose a cross dataset surrogate model called Differentiable Metafeature-based Surrogate (DMFBS), that predicts the hyperparameter response, i.e. validation loss, of a model trained on the dataset at hand. In contrast to existing models, DMFBS i) integrates a differentiable metafeature extractor and ii) is optimized using a novel multi-task loss, linking manifold regularization with a dataset similarity measure learned via an auxiliary dataset identification meta-task, effectively enforcing the response approximation for similar datasets to be similar. We compare DMFBS against several recent models for HPO on three large meta-datasets and show that it consistently outperforms all of them with an average 10% improvement. Finally, we provide an extensive ablation study that examines the different components of our approach.

Keywords

Cite

@article{arxiv.2102.03776,
  title  = {Hyperparameter Optimization with Differentiable Metafeatures},
  author = {Hadi S. Jomaa and Lars Schmidt-Thieme and Josif Grabocka},
  journal= {arXiv preprint arXiv:2102.03776},
  year   = {2021}
}
R2 v1 2026-06-23T22:54:42.938Z