Collaborative Online Learning of an Action Model
Résumé
A number of recent works have designed algorithms that allow an agent to revise a relational action model from interactions with its environment and uses this model for building plans and better exploring its environment. This article addresses Multi Agent Relational Action Learning: it considers a community of agents, each rationally acting following some relational action model, and assumes that the observed effect of past actions that led an agent to revise its action model can be communicated to other agents of the community, potentially speeding up the on-line learning process of agents in the community. We describe and experiment a framework for collaborative relational action model revision where each agent is autonomous and benefits from past observations memorized by all agents of the community.