A survey on multi-objective hyperparameter optimization algorithms for machine learning

Abstract Hyperparameter optimization (HPO) is a necessary step to ensure the best possible performance of Machine Learning (ML) algorithms. Several methods have been developed perform HPO; most these are focused on optimizing one measure (usually an error-based measure), and literature such single-objective HPO problems vast. Recently, though, algorithms appeared that focus multiple conflicting objectives simultaneously. This article presents systematic survey published between 2014 2020 multi-objective algorithms, distinguishing metaheuristic-based metamodel-based approaches using mixture both. We also discuss quality metrics used compare procedures present future research directions.