How to Boost Any Loss Function

Richard Nock, Yishay Mansour

Boosting is a highly successful ML-born optimization setting in which one is required to computationally efficiently learn arbitrarily good models based on the access to a weak lea…