Iterated Population Based Training with Task-Agnostic Restarts

arXiv (Cornell University)

Abstract

Hyperparameter Optimization (HPO) can lift the burden of tuning hyperparameters (HPs) of neural networks. HPO algorithms from the Population Based Training (PBT) family are efficient thanks to dynamically adjusting HPs every few steps of the weight optimization. Recent results indicate that the number of steps between HP updates is an important meta-HP of all PBT variants that can substantially affect their performance. Yet, no method or intuition is available for efficiently setting its value. We introduce Iterated Population Based Training (IPBT), a novel PBT variant that automatically adjusts this HP via restarts that reuse weight information in a task-agnostic way and leverage time-varying Bayesian optimization to reinitialize HPs. Evaluation on 8 image classification and reinforcement learning tasks shows that, on average, our algorithm matches or outperforms 5 previous PBT variants and other HPO algorithms (random search, ASHA, SMAC3), without requiring a budget increase or any changes to its HPs. The source code is available at https://github.com/AwesomeLemon/IPBT.

Authors 3

1. Alexander Chebykin

   Centrum Wiskunde & Informatica

   Affiliation as printed

   Centrum Wiskunde & Informatica , Amsterdam , the Netherlands

2. Tanja Alderliesten Aachen

   Leiden University Medical Center

   Affiliation as printed

   Leiden University Medical Center , Leiden , the Netherlands

3. Peter A. N. Bosman

   Delft University of Technology · Centrum Wiskunde & Informatica

   Affiliation as printed

   Centrum Wiskunde & Informatica , Amsterdam , the Netherlands

   Delft University of Technology , Delft , the Netherlands