Computable learning of natural hypothesis classes

Matthew Harrison-Trainor, Syed Akbari

This paper is about the recent notion of computably probably approximately correct learning, which lies between the statistical learning theory where there is no computational requ…