Understanding behaviors of a constructive memory agent: A markov chain analysis

摘要

This paper describes a Markov chain analysis of the behaviors of a constructive memory agent. It shows that a constructive memory agent behaves based on the knowledge structures that it has learned from its interaction with the environment. The internal reasoning moves from reactive and reflective to reflexive as the agent acquires more experiences that are increasingly grounded.