A0426
Title: Learning to defer to multiple experts: From individuals to populations and crowds
Authors: Eric Nalisnick - Johns Hopkins University (United States) [presenting]
Abstract: Artificial intelligence is being deployed in ever more consequential settings such as healthcare and autonomous driving. Thus, it is ensured that these systems are safe and trustworthy. One near-term solution is to ensure that a human is involved in the decision-making process and that the system can ask for help in difficult or high-risk scenarios. Recent advances are presented in the "learning to defer" paradigm: Decision-making responsibility is allocated to either a human or a model, depending on which is more likely to take the correction action. In particular, novel formulations are presented that can support multiple human decision makers, which could range from known individuals to anonymous members of a crowd.
