Multi-Objective Population Based Training

Arkadiy Dushatskiy; Alexander Chebykin; Tanja Alderliesten; Peter Bosman

ICML 2023poster5 citations

Multi-Objective Population Based Training

Arkadiy Dushatskiy, Alexander Chebykin, Tanja Alderliesten, Peter Bosman

Abstract

Population Based Training (PBT) is an efficient hyperparameter optimization algorithm. PBT is a single-objective algorithm, but many real-world hyperparameter optimization problems involve two or more conflicting objectives. In this work, we therefore introduce a multi-objective version of PBT, MO-PBT. Our experiments on diverse multi-objective hyperparameter optimization problems (Precision/Recall, Accuracy/Fairness, Accuracy/Adversarial Robustness) show that MO-PBT outperforms random search, single-objective PBT, and the state-of-the-art multi-objective hyperparameter optimization algorithm MO-ASHA.

BibTeX

@inproceedings{icml2023_multiobjectivepo,
  title = {Multi-Objective Population Based Training},
  author = {Arkadiy Dushatskiy and Alexander Chebykin and Tanja Alderliesten and Peter Bosman},
  booktitle = {ICML 2023},
  year = {2023}
}