paper-with-me

Papers

Model-Based Offline Reinforcement Learning with Pessimism-Modulated Dynamics Belief

2022-10-13 · Kaiyang Guo, Yunfeng Shao, Yanhui Geng

Model-based offline reinforcement learning (RL) aims to find highly rewarding policy, by leveraging a previously collected static dataset and a dynamics model. While the dynamics model learned through reuse of the static dataset, its generalization ability hopefully promotes policy learning if properly utilized. To that end, several works propose to quantify the uncertainty of predicted dynamics, and explicitly apply it to penalize reward. However, as the dynamics and the reward are intrinsically different factors in context of MDP, characterizing the impact of dynamics uncertainty through reward penalty may incur unexpected tradeoff between model utilization and risk avoidance. In this work, we instead maintain a belief distribution over dynamics, and evaluate/optimize policy through biased sampling from the belief. The sampling procedure, biased towards pessimism, is derived based on an alternating Markov game formulation of offline RL. We formally show that the biased sampling naturally induces an updated dynamics belief with policy-dependent reweighting factor, termed Pessimism-Modulated Dynamics Belief. To improve policy, we devise an iterative regularized policy optimization algorithm for the game, with guarantee of monotonous improvement under certain condition. To make practical, we further devise an offline RL algorithm to approximately find the solution. Empirical results show that the proposed approach achieves state-of-the-art performance on a wide range of benchmark tasks.

📄 PDF Abstract BibTeX arXiv:2210.06692

Code (3)

huawei-noah/HEBO/tree/master/PMDB 공식 구현 pytorch
huawei-noah/hebo 공식 구현 pytorch
2023-MindSpore-1/ms-code-220/tree/main/pmdb mindspore

Tasks

D4RLOffline RLreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Similar Papers 제목 키워드 기반

Value-Incentivized Preference Optimization: A Unified Approach to Online and Offline RLHF

2024-05-29 · Shicong Cen, Jincheng Mei, Katayoon Goshvadi, Hanjun Dai 외

Reinforcement learning from human feedback (RLHF) has demonstrated great promise in aligning large language models (LLMs) with human preference. Depending on the availability of preference data, both online and offline R…

reinforcement-learningReinforcement LearningReinforcement Learning (RL)Text Summarization

State-Aware Proximal Pessimistic Algorithms for Offline Reinforcement Learning

2022-11-28 · Chen Chen, Hongyao Tang, Yi Ma, Chao Wang 외

Pessimism is of great importance in offline reinforcement learning (RL). One broad category of offline RL algorithms fulfills pessimism by explicit or implicit behavior regularization. However, most of them only consider…

Offline RLQ-Learningreinforcement-learningReinforcement Learning+1

Model-based Offline Reinforcement Learning with Local Misspecification

2023-01-26 · Kefan Dong, Yannis Flet-Berliac, Allen Nie, Emma Brunskill

We present a model-based offline reinforcement learning policy performance lower bound that explicitly captures dynamics model misspecification and distribution mismatch and we propose an empirical algorithm for optimal …

D4RLmodelreinforcement-learningReinforcement Learning+1

Distorted Distributional Policy Evaluation for Offline Reinforcement Learning

2026-01-05 · Ryo Iwaki, Takayuki Osogami arxiv

While Distributional Reinforcement Learning (DRL) methods have demonstrated strong performance in online settings, its success in offline scenarios remains limited. We hypothesize that a key limitation of existing offlin…

Reinforcement Learning

Bellman-consistent Pessimism for Offline Reinforcement Learning

2021-06-13 · NeurIPS 2021 12 · Tengyang Xie, Ching-An Cheng, Nan Jiang, Paul Mineiro 외

The use of pessimism, when reasoning about datasets lacking exhaustive exploration has recently gained prominence in offline reinforcement learning. Despite the robustness it adds to the algorithm, overly pessimistic rea…

reinforcement-learningReinforcement LearningReinforcement Learning (RL)