A Preference-based Interpretation of Other Agents’ Actions

Jérôme Lang

What can we infer about the beliefs, the goals and the intended next actions of an agent by observing him acting? We start from the same intuitive principle as in (Brafman and Tennenholtz 1997), namely, that an agent normally follows some optimal plan (with respect to some preference relation) and chooses actions that are part of this plan. This principle allows us to infer beliefs about the agent’s plausible next actions, about his goals and about his initial and current beliefs. We give several possible definitions and study their respective meanings, especially with respect to the assumptions about the acting agent’s rationality.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.