Majority-of-Three: The Simplest Optimal Learner?

arXiv:2403.08831v1 Announce Type: cross
Abstract: Developing an optimal PAC learning algorithm in the realizable setting, where empirical risk minimization (ERM) is suboptimal, was a major open problem in learning theory for decades. The problem was finally resolved by Hanneke a few years ago. Unfortunately, Hanneke's algorithm is quite complex as it returns the majority vote of many ERM classifiers that are trained on carefully selected subsets of the data. It is thus a natural goal to determine the simplest algorithm …

abstract algorithm arxiv cs.lg erm finally major math.st returns risk stat.ml stat.th theory type vote