Learning to Make Adherence-Aware Advice
This work addresses the challenge of AI systems providing advice that humans may ignore, which is an incremental improvement for applications involving AI-assisted decision-making.
The paper tackles the problem of suboptimal AI policies in human-AI interactions by developing a sequential decision-making model that accounts for human adherence levels and includes a defer option for selective advice. The result is specialized learning algorithms that outperform general reinforcement learning methods in both theoretical convergence and empirical performance.
As artificial intelligence (AI) systems play an increasingly prominent role in human decision-making, challenges surface in the realm of human-AI interactions. One challenge arises from the suboptimal AI policies due to the inadequate consideration of humans disregarding AI recommendations, as well as the need for AI to provide advice selectively when it is most pertinent. This paper presents a sequential decision-making model that (i) takes into account the human's adherence level (the probability that the human follows/rejects machine advice) and (ii) incorporates a defer option so that the machine can temporarily refrain from making advice. We provide learning algorithms that learn the optimal advice policy and make advice only at critical time stamps. Compared to problem-agnostic reinforcement learning algorithms, our specialized learning algorithms not only enjoy better theoretical convergence properties but also show strong empirical performance.