English

Meta-Learning for Symbolic Hyperparameter Defaults

Machine Learning 2021-10-07 v2 Machine Learning

Abstract

Hyperparameter optimization in machine learning (ML) deals with the problem of empirically learning an optimal algorithm configuration from data, usually formulated as a black-box optimization problem. In this work, we propose a zero-shot method to meta-learn symbolic default hyperparameter configurations that are expressed in terms of the properties of the dataset. This enables a much faster, but still data-dependent, configuration of the ML algorithm, compared to standard hyperparameter optimization approaches. In the past, symbolic and static default values have usually been obtained as hand-crafted heuristics. We propose an approach of learning such symbolic configurations as formulas of dataset properties from a large set of prior evaluations on multiple datasets by optimizing over a grammar of expressions using an evolutionary algorithm. We evaluate our method on surrogate empirical performance models as well as on real data across 6 ML algorithms on more than 100 datasets and demonstrate that our method indeed finds viable symbolic defaults.

Keywords

Cite

@article{arxiv.2106.05767,
  title  = {Meta-Learning for Symbolic Hyperparameter Defaults},
  author = {Pieter Gijsbers and Florian Pfisterer and Jan N. van Rijn and Bernd Bischl and Joaquin Vanschoren},
  journal= {arXiv preprint arXiv:2106.05767},
  year   = {2021}
}

Comments

Pieter Gijsbers and Florian Pfisterer contributed equally to the paper. V1: Two page GECCO poster paper accepted at GECCO 2021. V2: The original full length paper (8 pages) with appendix