2022/07/30 by Kelly Zhang, Kelly W. Zhang, Omer Gottesman +5
Computer Science · Decision Sciences · Mathematics · #Advanced Bandit Algorithms Research #Artificial intelligence #Bayesian probability #Computer science #Data Stream Mining Techniques #Domain (mathematical analysis) #Machine learning #Markov chain #Markov decision process #Markov process #Mathematical optimization #Mathematics #Multi-armed bandit #Regret #Reinforcement Learning in Robotics #Reinforcement learning #cs.AI #cs.LG
paper · pdf · doi:10.48550/arxiv.2208.00250
published in arXiv (Cornell University) (Cornell University) · Challenges of Real-World Reinforcement Learning 2020 (NeurIPS Workshop)
arxiv created 2022/07/30 · openalex publication_date 2022/07/30 · arxiv updated 2022/08/02 · openalex created_date 2022/08/03 · openalex updated_date 2026/07/28
In the reinforcement learning literature, there are many algorithms developed for either Contextual Bandit (CB) or Markov Decision Processes (MDP) environments. However, when deploying reinforcement learning algorithms in the real world, even with domain expertise, it is often difficult to know whether it is appropriate to treat a sequential decision making problem as a CB or an MDP. In other words, do actions affect future states, or only the immediate rewards? Making the wrong assumption regarding the nature of the environment can lead to inefficient learning, or even prevent the algorithm from ever learning an optimal policy, even with infinite data. In this work we develop an online algorithm that uses a Bayesian hypothesis testing approach to learn the nature of the environment. Our algorithm allows practitioners to incorporate prior knowledge about whether the environment is that of a CB or an MDP, and effectively interpolate between classical CB and MDP-based algorithms to mitigate against the effects of misspecifying the environment. We perform simulations and demonstrate that in CB settings our algorithm achieves lower regret than MDP-based algorithms, while in non-bandit MDP settings our algorithm is able to learn the optimal policy, often achieving comparable regret to MDP-based algorithms.