CS 598 : Theoretical Machine Learning


In the previous class, we had looked at the RMW algorithm. In this setting, the actions/decisions are referred to as experts. Assume that we have N experts. On any given day t, we choose an expert to use from among the N experts (like choosing a function h ∈ H). The adversary/world returns a loss vector lt = (lt 1, · · · lt N ). If the expert it is chosen… (More)


2 Figures and Tables

Slides referencing similar topics