Effective Adaptive Mutation Rates for Program Synthesis

Created by W.Langdon from gp-bibliography.bib Revision:1.8120

@InProceedings{ni:2024:GECCO,

• author = "Andrew Ni and Lee Spector",
• title = "Effective Adaptive Mutation Rates for Program Synthesis",
• booktitle = "Proceedings of the 2024 Genetic and Evolutionary Computation Conference",
• year = "2024",
• editor = "Ting Hu and Aniko Ekart and Julia Handl and Xiaodong Li and Markus Wagner and Mario Garza-Fabre and Kate Smith-Miles and Richard Allmendinger and Ying Bi and Grant Dick and Amir H Gandomi and Marcella Scoczynski Ribeiro Martins and Hirad Assimi and Nadarajen Veerapen and Yuan Sun and Mario Andres Munyoz and Ahmed Kheiri and Nguyen Su and Dhananjay Thiruvady and Andy Song and Frank Neumann and Carla Silva",
• pages = "952--960",
• address = "Melbourne, Australia",
• series = "GECCO '24",
• month = "14-18 " # jul,
• organisation = "SIGEVO",
• publisher = "Association for Computing Machinery",
• publisher_address = "New York, NY, USA",
• keywords = "genetic algorithms, genetic programming, uniform addition by mutation and deletion, adaptive mutation rate, max K-armed bandits, software synthesis",
• isbn13 = "979-8-4007-0494-9",
• DOI = "doi:10.1145/3638529.3654135",
• size = "9 pages",
• abstract = "The problem-solving performance of many evolutionary algorithms, including genetic programming systems used for program synthesis, depends on the values of hyperparameters including mutation rates. The mutation method used to produce some of the best results to date on software synthesis benchmark problems, Uniform Mutation by Addition and Deletion (UMAD), adds new genes into a genome at a predetermined rate and then deletes genes at a rate that balances the addition rate, producing no size change on average. While UMAD with a predetermined addition rate outperforms many other mutation and crossover schemes, we do not expect a single rate to be optimal across all problems or all generations within one run of an evolutionary system. However, many current adaptive mutation schemes such as self-adaptive mutation rates suffer from pathologies like the vanishing mutation rate problem, in which the mutation rate quickly decays to zero. We propose an adaptive bandit-based scheme that addresses this problem and essentially removes the need to specify a mutation rate. Although the proposed scheme itself introduces hyperparameters, we either set these to good values or ensemble them in a reasonable range. Results on software synthesis and symbolic regression problems validate the effectiveness of our approach.",
• notes = "GECCO-2024 GP A Recombination of the 33rd International Conference on Genetic Algorithms (ICGA) and the 29th Annual Genetic Programming Conference (GP)",

}

Genetic Programming entries for Andrew Ni Lee Spector

Citations