paper-with-me

Papers

Deterministic Uncertainty Propagation for Improved Model-Based Offline Reinforcement Learning

2024-06-06 · Abdullah Akgül, Manuel Haußmann, Melih Kandemir

Current approaches to model-based offline reinforcement learning often incorporate uncertainty-based reward penalization to address the distributional shift problem. These approaches, commonly known as pessimistic value iteration, use Monte Carlo sampling to estimate the Bellman target to perform temporal difference-based policy evaluation. We find out that the randomness caused by this sampling step significantly delays convergence. We present a theoretical result demonstrating the strong dependency of suboptimality on the number of Monte Carlo samples taken per Bellman target calculation. Our main contribution is a deterministic approximation to the Bellman target that uses progressive moment matching, a method developed originally for deterministic variational inference. The resulting algorithm, which we call Moment Matching Offline Model-Based Policy Optimization (MOMBO), propagates the uncertainty of the next state through a nonlinear Q-network in a deterministic fashion by approximating the distributions of hidden layer activations by a normal distribution. We show that it is possible to provide tighter guarantees for the suboptimality of MOMBO than the existing Monte Carlo sampling approaches. We also observe MOMBO to converge faster than these approaches in a large set of benchmark tasks.

📄 PDF Abstract BibTeX arXiv:2406.04088

Code (1)

adinlab/MOMBO 공식 구현 pytorch

Tasks

reinforcement-learningReinforcement LearningReinforcement Learning (RL)Variational Inference

Methods 이 논문이 사용한 방법론

SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

One Risk to Rule Them All: A Risk-Sensitive Perspective on Model-Based Offline Reinforcement Learning

2022-11-30 · NeurIPS 2023 11 · Marc Rigter, Bruno Lacerda, Nick Hawes

Offline reinforcement learning (RL) is suitable for safety-critical domains where online exploration is too costly or dangerous. In such safety-critical settings, decision-making should take into consideration the risk o…

AllDecision MakingOffline RLReinforcement Learning (RL)+1

Behavior-Adaptive Q-Learning: A Unifying Framework for Offline-to-Online RL

2025-11-05 · Lipeng Zu, Hansong Zhou, Xiaonan Zhang arxiv

Offline reinforcement learning (RL) enables training from fixed data without online interaction, but policies learned offline often struggle when deployed in dynamic environments due to distributional shift and unreliabl…

Reinforcement Learning

Selective Uncertainty Propagation in Offline RL

2023-02-01 · Sanath Kumar Krishnamurthy, Tanmay Gangwani, Sumeet Katariya, Branislav Kveton 외

We consider the finite-horizon offline reinforcement learning (RL) setting, and are motivated by the challenge of learning the policy at any step h in dynamic programming (DP) algorithms. To learn this, it is sufficient …

Offline RLreinforcement-learningReinforcement Learning (RL)

Control invariant set enhanced safe reinforcement learning: improved sampling efficiency, guaranteed stability and robustness

2023-05-24 · Song Bo, Bernard T. Agyeman, Xunyuan Yin, Jinfeng Liu

Reinforcement learning (RL) is an area of significant research interest, and safe RL in particular is attracting attention due to its ability to handle safety-driven constraints that are crucial for real-world applicatio…

Reinforcement Learning (RL)Safe Reinforcement Learning

Epistemic Robust Offline Reinforcement Learning

2026-04-08 · Abhilash Reddy Chenreddy, Erick Delage arxiv

Offline reinforcement learning learns policies from fixed datasets without further environment interaction. A key challenge in this setting is epistemic uncertainty, arising from limited or biased data coverage, particul…

Reinforcement LearningOffline RL