Discovering Attention-Based Genetic Algorithms via Meta-Black-Box Optimization
Open Access
- 12 July 2023
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 929-937
Abstract
Genetic algorithms constitute a family of black-box optimization algorithms, which take inspiration from the principles of biological evolution. While they provide a general-purpose tool for optimization, their particular instantiations can be heuristic and motivated by loose biological intuition. In this work we explore a fundamentally different approach: Given a sufficiently flexible parametrization of the genetic operators, we discover entirely new genetic algorithms in a data-driven fashion. More specifically, we parametrize selection and mutation rate adaptation as cross- and self-attention modules and use Meta-Black-Box-Optimization to evolve their parameters on a set of diverse optimization tasks. The resulting Learned Genetic Algorithm outperforms state-of-the-art adaptive baseline genetic algorithms and generalizes far beyond its meta-training settings. The learned algorithm can be applied to previously unseen optimization problems, search dimensions & evaluation budgets. We conduct extensive analysis of the discovered operators and provide ablation experiments, which highlight the benefits of flexible module parametrization and the ability to transfer ('plug-in') the learned operators to conventional genetic algorithms.Keywords
This publication has 8 references indexed in Scilit:
- Introducing Symmetries to Black Box Meta Reinforcement LearningProceedings of the AAAI Conference on Artificial Intelligence, 2022
- Learning Not to Learn: Nature versus Nurture In SilicoProceedings of the AAAI Conference on Artificial Intelligence, 2022
- Automated Reinforcement Learning (AutoRL): A Survey and Open ProblemsJournal of Artificial Intelligence Research, 2022
- seaborn: statistical data visualizationJournal of Open Source Software, 2021
- Learning Step-Size Adaptation in CMA-ES2020
- Quality and Diversity Optimization: A Unifying Modular FrameworkIEEE Transactions on Evolutionary Computation, 2017
- Natural Selection Fails to Optimize Mutation Rates for Long-Term Adaptation on Rugged Fitness LandscapesPLOS Computational Biology, 2008
- A Simple Modification in CMA-ES Achieving Linear Time and Space Complexity2008