ABSTRACT
We present an analysis of the loss of population-level test coverage induced by different down-sampling strategies when combined with lexicase selection. We study recorded populations from the first generation of genetic programming runs, as well as entirely synthetic populations. Our findings verify the hypothesis that informed down-sampling better maintains population-level test coverage when compared to random down-sampling. Additionally, we show that both forms of down-sampling cause greater test coverage loss than standard lexicase selection with no down-sampling. However, given more information about the population, we found that informed down-sampling can further reduce its test coverage loss. We also recommend wider adoption of the static population analyses we present in this work.
- Ryan Boldi, Ashley Bao, Martin Briesch, Thomas Helmuth, Dominik Sobania, Lee Spector, and Alexander Lalejini. 2023. The Problem Solving Benefits of Down-sampling Vary by Selection Scheme. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (Lisbon, Portugal) (GECCO '23). Google ScholarDigital Library
- Ryan Boldi, Martin Briesch, Dominik Sobania, Alexander Lalejini, Thomas Helmuth, Franz Rothlauf, Charles Ofria, and Lee Spector. 2023. Informed Down-Sampled Lexicase Selection: Identifying productive training cases for efficient problem solving. arXiv:2301.01488. Google ScholarCross Ref
- Li Ding and Lee Spector. 2021. Optimizing neural networks with gradient lexicase selection. In International Conference on Learning Representations.Google Scholar
- Thomas Helmuth and Peter Kelly. 2021. PSB2: The Second Program Synthesis Benchmark Suite. arXiv:2106.06086. Google ScholarCross Ref
- Thomas Helmuth and Peter Kelly. 2022. Applying genetic programming to PSB2: the next generation program synthesis benchmark suite. Genetic Programming and Evolvable Machines (June 2022). Google ScholarDigital Library
- Thomas Helmuth and Lee Spector. 2015. General Program Synthesis Benchmark Suite. In Proceedings of the 2015 Annual Conference on Genetic and Evolutionary Computation. ACM, Madrid Spain, 1039--1046. Google ScholarDigital Library
- Thomas Helmuth and Lee Spector. 2021. Problem-solving benefits of down-sampled lexicase selection. Artificial Life (jun 2021), 1--21. arXiv:2106.06085 Google ScholarCross Ref
- Thomas Helmuth, Lee Spector, and James Matheson. 2015. Solving Uncompromising Problems With Lexicase Selection. IEEE Transactions on Evolutionary Computation 19, 5 (2015), 630--643. Google ScholarDigital Library
- Jose Guadalupe Hernandez, Alexander Lalejini, Emily Dolson, and Charles Ofria. 2019. Random subsampling improves performance in lexicase selection. In Proceedings of the Genetic and Evolutionary Computation Conference Companion. ACM, Prague Czech Republic, 2028--2031. Google ScholarDigital Library
- Jose Guadalupe Hernandez, Alexander Lalejini, and Charles Ofria. 2022. An Exploration of Exploration: Measuring the Ability of Lexicase Selection to Find Obscure Pathways to Optimality. In Genetic Programming Theory and Practice XVIII, Wolfgang Banzhaf, Leonardo Trujillo, Stephan Winkler, and Bill Worzel (Eds.). Springer Nature Singapore, Singapore, 83--107. Google ScholarCross Ref
- Alexander Lalejini, Emily Dolson, Anya E Vostinar, and Luis Zaman. 2022. Artificial selection methods from evolutionary computing show promise for directed evolution of microbes. eLife 11 (Aug. 2022), e79665. Google ScholarCross Ref
- Christian W.G. Lasarczyk, Peter Dittrich, and Wolfgang Banzhaf. 2004. Dynamic Subset Selection Based on a Fitness Case Topology. Evolutionary Computation 12, 2 (June 2004), 223--242. Google ScholarDigital Library
- Jared M. Moore and Adam Stanton. 2017. Lexicase selection outperforms previous strategies for incremental evolution of virtual creature controllers. In Proceedings of the Fourteenth European Conference Artificial Life, ECAL 2017, Lyon, France, September 4--8, 2017. MIT Press, 290--297. http://cognet.mit.edu/journal/ecal2017Google Scholar
- Lee Spector. 2012. Assessment of Problem Modality by Differential Performance of Lexicase Selection in Genetic Programming: A Preliminary Report. In Proceedings of the 14th Annual Conference Companion on Genetic and Evolutionary Computation (Philadelphia, Pennsylvania, USA) (GECCO '12). Association for Computing Machinery, New York, NY, USA, 401--408. Google ScholarDigital Library
- Adam Stanton and Jared M. Moore. 2022. Lexicase Selection for Multi-Task Evolutionary Robotics. Artificial Life 28, 4 (Nov. 2022), 479--498. Google ScholarCross Ref
Recommendations
The Problem Solving Benefits of Down-sampling Vary by Selection Scheme
GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary ComputationGenetic programming systems often use large training sets to evaluate candidate solutions, which can be computationally expensive. Down-sampling training sets has long been used to decrease the computational cost of evaluation in a wide range of ...
Epsilon-Lexicase Selection for Regression
GECCO '16: Proceedings of the Genetic and Evolutionary Computation Conference 2016Lexicase selection is a parent selection method that considers test cases separately, rather than in aggregate, when performing parent selection. It performs well in discrete error spaces but not on the continuous-valued problems that compose most ...
An Adaptive Image Resizing Algorithm in DCT Domain
A novel image resizing algorithm is proposed. In our method, three steps are included in the downsampling: the first-round downsampling, the interim upsampling and the second-round downsampling. The downsampling operation unit size is selected between ...
Comments