Best-Response Bayesian Reinforcement Learning with Bayes-adaptive POMDPs for Centaurs
This addresses the challenge of human-AI collaboration in partially observable tasks for centaurs, though it appears incremental as it builds on existing POMDP and game theory frameworks.
The paper tackles the problem of AI assisting bounded-rational humans in decision-making by formulating their interaction as a sequential game with Bayesian best-response models, reducing it to a Bayes-adaptive POMDP, and shows in simulations that the AI can infer human constraints and nudge them towards better decisions.
Centaurs are half-human, half-AI decision-makers where the AI's goal is to complement the human. To do so, the AI must be able to recognize the goals and constraints of the human and have the means to help them. We present a novel formulation of the interaction between the human and the AI as a sequential game where the agents are modelled using Bayesian best-response models. We show that in this case the AI's problem of helping bounded-rational humans make better decisions reduces to a Bayes-adaptive POMDP. In our simulated experiments, we consider an instantiation of our framework for humans who are subjectively optimistic about the AI's future behaviour. Our results show that when equipped with a model of the human, the AI can infer the human's bounds and nudge them towards better decisions. We discuss ways in which the machine can learn to improve upon its own limitations as well with the help of the human. We identify a novel trade-off for centaurs in partially observable tasks: for the AI's actions to be acceptable to the human, the machine must make sure their beliefs are sufficiently aligned, but aligning beliefs might be costly. We present a preliminary theoretical analysis of this trade-off and its dependence on task structure.