Related papers

[folivetti]

This paper is somehow related to the Brush project

https://www.researchgate.net/publication/369615357_Domain-Aware_Feature_Learning_with_Grammar-Guided_Genetic_Programming

The main idea is to constrain the set of operators depending on the type of the feature being processed at that branch. So, for example, if it is a categorical feature, it will only compare to equality with one of the allowed values. It does seem to be an early stage, but possibly it will intersect with brush eventually.

This one is my paper at gecco, even though it is about program synthesis, but it also implements type checks to constraint the program tree generation:

https://arxiv.org/abs/2304.03200