Responsible AI (RAI) Games and Ensembles
This work provides a general framework for addressing societal AI issues like fairness and robustness, but it is incremental as it builds on existing min-max formulations.
The authors tackled the problem of minimizing worst-case loss over uncertainty sets in AI objectives like fairness and robustness by proposing a general framework called Responsible AI (RAI) games, with results showing competitive performance in empirical tests, particularly for subpopulation shift.
Several recent works have studied the societal effects of AI; these include issues such as fairness, robustness, and safety. In many of these objectives, a learner seeks to minimize its worst-case loss over a set of predefined distributions (known as uncertainty sets), with usual examples being perturbed versions of the empirical distribution. In other words, aforementioned problems can be written as min-max problems over these uncertainty sets. In this work, we provide a general framework for studying these problems, which we refer to as Responsible AI (RAI) games. We provide two classes of algorithms for solving these games: (a) game-play based algorithms, and (b) greedy stagewise estimation algorithms. The former class is motivated by online learning and game theory, whereas the latter class is motivated by the classical statistical literature on boosting, and regression. We empirically demonstrate the applicability and competitive performance of our techniques for solving several RAI problems, particularly around subpopulation shift.