Pitfalls and Best Practices in Algorithm Configuration

authored by

Katharina Eggensperger, Marius Lindauer, Frank Hutter

Abstract

Good parameter settings are crucial to achieve high performance in many areas of artificial intelligence (AI), such as propositional satisfiability solving, AI planning, scheduling, and machine learning (in particular deep learning). Automated algorithm configuration methods have recently received much attention in the AI community since they replace tedious, irreproducible and error-prone manual parameter tuning and can lead to new state-of-the-art performance. However, practical applications of algorithm configuration are prone to several (often subtle) pitfalls in the experimental design that can render the procedure ineffective. We identify several common issues and propose best practices for avoiding them. As one possibility for automatically handling as many of these as possible, we also propose a tool called GenericWrapper4AC.

External Organisation(s)

University of Freiburg

Type

Article

Journal

Journal of Artificial Intelligence Research

Volume

64

Pages

861-893

No. of pages

33

ISSN

1076-9757

Publication date

16.04.2019

Publication status

Published

Peer reviewed

Yes

ASJC Scopus subject areas

Artificial Intelligence

Electronic version(s)

https://arxiv.org/abs/1705.06058v3 (Access: Open)
https://doi.org/10.1613/jair.1.11420 (Access: Open)