What Do You Think I Think? Accounting for Human Beliefs Using Second-Order Theory of Mind
For human-agent interaction, this work addresses the problem of belief discrepancies caused by cognitive biases, offering a method to improve communication and learning.
This work uses second-order Theory of Mind (ToM-2) to enable agents to detect and account for human cognitive biases and heuristics during interactions. In a user study, the ToM-2 learner significantly improved the informativeness of teacher actions and received higher subjective usefulness ratings.
Discrepancies between an agent's actual knowledge and what a person thinks the agent knows can hinder interactions. If an agent could detect such discrepancies, it could provide feedback to account for them and improve current and future interactions. Using the I-POMDP as a framework for a second-order Theory of Mind (ToM-2), this work endows an agent with the ability to model the evolution of a person's erroneous beliefs about an agent and the cognitive biases and heuristics (CBH) from which they arise. In doing so, the agent can detect when CBH might be at play during an interaction and adaptively generate feedback that accounts for them. An in-person user study shows how a ToM-2 learner can account for the effects of a teacher's CBH to significantly improve the informativeness of teacher actions, and subjective results suggest people find the ToM-2 learner's feedback more useful.