Scaling Laws for Hyperparameter Optimization

Arlind Kadra, Maciej Janowski, Martin Wistuba, Josif Grabocka

Keywords: hyperparameter optimization, multi-fidelity hyperparameter optimization, multi-fidelity hpo, power laws, deep neural networks, deep power laws, deep ensemble, deep learning, large language models, scaling laws, llm

TL;DR: Multi-fidelity hyperparameter optimization with deep power laws that achieves state-of-the-art results across diverse benchmarks.

Abstract: Hyperparameter optimization is an important subfield of machine learning that focuses on tuning the hyperparameters of a chosen algorithm to achieve peak performance. Recently, there has been a stream of methods that tackle the issue of hyperparameter optimization, however, most of the methods do not exploit the dominant power law nature of learning curves for Bayesian optimization. In this work, we propose Deep Power Laws (DPL), an ensemble of neural network models conditioned to yield predictions that follow a power-law scaling pattern. Our method dynamically decides which configurations to pause and train incrementally by making use of gray-box evaluations. We compare our method against 7 state-of-the-art competitors on 3 benchmarks related to tabular, image, and NLP datasets covering 59 diverse tasks. Our method achieves the best results across all benchmarks by obtaining the best any-time results compared to all competitors.

Submission Number: 695