Learning When to Take Advice: A Statistical Test for Achieving A Correlated Equilibrium
This addresses the challenge of trust and efficiency in multiagent systems by providing a statistical test for advice-taking, though it appears incremental as it builds on existing learning frameworks.
The paper tackles the problem of multiagent learning where agents must decide whether to follow a mediator's advice, presenting an algorithm that verifies the usefulness of the advice with high probability, enabling agents to reach a correlated equilibrium if the advice is useful or fall back to their original learning algorithm if not.
We study a multiagent learning problem where agents can either learn via repeated interactions, or can follow the advice of a mediator who suggests possible actions to take. We present an algorithmthat each agent can use so that, with high probability, they can verify whether or not the mediator's advice is useful. In particular, if the mediator's advice is useful then agents will reach a correlated equilibrium, but if the mediator's advice is not useful, then agents are not harmed by using our test, and can fall back to their original learning algorithm. We then generalize our algorithm and show that in the limit it always correctly verifies the mediator's advice.