Combinatorial learning theory

A theoretical direction that studies how to understand and design learning algorithms from the perspective of combinatorial structures, focusing on the relationship between complexity measures of sequences and structures and their learnability.