Pitfalls and Best Practices in Algorithm Configuration
Preprint
Abstract
Good parameter settings are crucial to achieve high performance in many areas of artificial intelligence (AI), such as 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, including a tool called GenericWrapper4AC for preventing the many possible problems in measuring the performance of the algorithm being optimized by executing it in a standardized, controlled manner.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- MaxSAT by improved instance-specific algorithm configurationArtificial Intelligence, 2016
- AutoFolio: An Automatically Configured Algorithm SelectorJournal of Artificial Intelligence Research, 2015
- Automatic Component-Wise Design of Multiobjective Evolutionary AlgorithmsIEEE Transactions on Evolutionary Computation, 2015
- Algorithm Selection for Combinatorial Search Problems: A SurveyAI Magazine, 2014
- Bridging the gap between easy generation and efficient verification of unsatisfiability proofsSoftware Testing, Verification and Reliability, 2014
- Algorithm runtime prediction: Methods & evaluationArtificial Intelligence, 2014
- Conflict-driven answer set solving: From theory to practiceArtificial Intelligence, 2012
- Controlling a Solver Execution with the runsolver ToolJournal on Satisfiability, Boolean Modeling and Computation, 2011
- Tradeoffs in the empirical evaluation of competing algorithm designsAnnals of Mathematics and Artificial Intelligence, 2010
- ParamILS: An Automatic Algorithm Configuration FrameworkJournal of Artificial Intelligence Research, 2009