Exploring Environmental Change for Down-Sampled Lexicase Selection

Anonymous

Keywords: Environmental Change, Genetic Programming, Lexicase Selection, Down-sampling

TL;DR: This paper proposes and studies two varieties of down-sampled lexicase selection for GP that affect the speed and magnitude of environmental change.

Abstract: Down-sampling training data has long been shown to improve the generalization performance of a wide range of machine learning systems. Recently, down-sampling has proved effective in genetic programming (GP) runs that utilize the lexicase parent selection technique. Although this down-sampling procedure has been shown to significantly improve performance across a variety of problems, it does not seem to do so due to encouraging adaptability through environmental change. We investigate modifications to down-sampled lexicase selection in hopes of promoting incremental environmental change to scaffold evolution by reducing the amount of jarring discontinuities between the environments of successive generations. In our empirical studies, we find that promoting environmental change by rolling the down-sample across the entire training set is not significantly better for evolving solutions to program synthesis problems than simple random down-sampling. We also find that going in the other direction by using only disjoint down-samples also does not significantly differ from the performance of regular random down-sampling. These results highlight some new insights into the role of environmental change for down-sampled lexicase selection, and present a viable new direction for future research.