Bayes-adaptive Deep Model-based Policy Optimisation

Abstract

We introduce a Bayesian (deep) model-based reinforcement learning method (RoMBRL) that can capture model uncertainty to achieve sample-efficient policy optimisation. We propose to formulate the model-based policy optimisation problem as a Bayes-adaptive Markov decision process (BAMDP). RoMBRL maintains model uncertainty via belief distributions through a deep Bayesian neural network whose samples are generated via stochastic gradient Hamiltonian Monte Carlo. Uncertainty is propagated through simulations controlled by sampled models and history-based policies. As beliefs are encoded in visited histories, we propose a history-based policy network that can be end-to-end trained to generalise across history space and will be trained using recurrent Trust-Region Policy Optimisation. We show that RoMBRL outperforms existing approaches on many challenging control benchmark tasks in terms of sample complexity and task performance. The source code of this paper is also publicly available on htt

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyhoang2020bayes

Related papers

• Bayesian Policy Optimization For Model Uncertainty (2018)0.00
• Enhancing Offline Model-based RL Via Active Model Selection: A Bayesian Optimization Perspective (2025)0.00
• Deep Model-based Reinforcement Learning Via Estimated Uncertainty And Conservative Policy Optimization (2019)0.00
• Conservative Dual Policy Optimization For Efficient Model-based Reinforcement Learning (2022)0.00
• Bayes Adaptive Monte Carlo Tree Search For Offline Model-based Reinforcement Learning (2024)0.00
• Bayesian Residual Policy Optimization: Scalable Bayesian Reinforcement Learning With Clairvoyant Experts (2020)0.00
• TOM: Learning Policy-aware Models For Model-based Reinforcement Learning Via Transition Occupancy Matching (2023)0.00
• Double Horizon Model-based Policy Optimization (2025)0.00