Dynamic Knowledge Injection for AIXI Agents
This addresses a fundamental limitation in reinforcement learning for human-AI teaming, though it is incremental as it builds on existing AIXI approximations.
The paper tackles the problem of epistemic uncertainty in AIXI approximations by introducing DynamicHedgeAIXI, an agent that dynamically incorporates new models from a human operator, achieving good performance guarantees and validated utility in epidemic control experiments.
Prior approximations of AIXI, a Bayesian optimality notion for general reinforcement learning, can only approximate AIXI's Bayesian environment model using an a-priori defined set of models. This is a fundamental source of epistemic uncertainty for the agent in settings where the existence of systematic bias in the predefined model class cannot be resolved by simply collecting more data from the environment. We address this issue in the context of Human-AI teaming by considering a setup where additional knowledge for the agent in the form of new candidate models arrives from a human operator in an online fashion. We introduce a new agent called DynamicHedgeAIXI that maintains an exact Bayesian mixture over dynamically changing sets of models via a time-adaptive prior constructed from a variant of the Hedge algorithm. The DynamicHedgeAIXI agent is the richest direct approximation of AIXI known to date and comes with good performance guarantees. Experimental results on epidemic control on contact networks validates the agent's practical utility.