Is One Step Enough for Offline Policy Improvement?

arXiv:2609.03842 · cs.LG · Submitted 2026-09-03 · Read on arXiv

cs.LG

Submitted: 2026-09-03

Updated: 2026-09-29

Comments: Preprint; 22 pages. Major revision with a new title, revised analysis of horizon subdivision and re-centering, and expanded experiments and controls

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Actor — 45× in this paper · explained in 3 episodes
• Reinforcement Learning — 24× in this paper · explained in 4 episodes
• Gradient Flow — 14× in this paper · explained in 2 episodes
• Q-learning — 6× in this paper · explained in 3 episodes
• Distribution Shift — 5× in this paper · explained in 5 episodes
• Optimal Transport — 5× in this paper · explained in 5 episodes
• Benchmark — 4× in this paper · explained in 2 episodes
• Regularization — 4× in this paper · explained in 3 episodes
• Transformer — 3× in this paper · explained in 2 episodes
• Action Space — 2× in this paper · explained in 1 episode
• Deep Reinforcement Learning — 2× in this paper · explained in 2 episodes
• Reinforcement Learning (RL) — 2× in this paper · explained in 25 episodes
• Wasserstein Distance — 2× in this paper · explained in 4 episodes
• Ablation — 1× in this paper · explained in 1 episode
• Diffusion Model — 1× in this paper · explained in 8 episodes
• Distribution Matching — 1× in this paper · explained in 2 episodes
• Fisher Information — 1× in this paper · explained in 3 episodes
• Markov Decision Process (MDP) — 1× in this paper · explained in 11 episodes
• Online Reinforcement Learning — 1× in this paper · explained in 2 episodes
• Optimal Transport (OT) — 1× in this paper · explained in 4 episodes
• Reward Function — 1× in this paper · explained in 2 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Trace — 1× in this paper · explained in 2 episodes

Sources

• Q-Transformer: Scalable Offline Reinforcement Learning via Autoregressive Q-Functions
• Optimal transport natural gradient for statistical manifolds with continuous sample space
• FlowQ: Energy-Guided Flow Policies for Offline Reinforcement Learning
• Uncertainty-Based Offline Reinforcement Learning with Diversified Q-Ensemble
• Sinkhorn Distances: Lightspeed Computation of Optimal Transportation Distances
• Forward-backward Gaussian variational inference via JKO in the Bures-Wasserstein Space
• Off-Policy Deep Reinforcement Learning without Exploration
• A Brief Introduction to Manifold Optimization
• Offline Reinforcement Learning with Implicit Q-Learning
• Stabilizing Off-Policy Q-Learning via Bootstrapping Error Reduction
• Conservative Q-Learning for Offline Reinforcement Learning
• PPO in the Fisher-Rao geometry
• Q-learning with Adjoint Matching
• Efficient Wasserstein Natural Gradients for Reinforcement Learning
• AWAC: Accelerating Online Reinforcement Learning with Offline Datasets
• Flow Q-Learning
• Revisiting the Minimalist Approach to Offline Reinforcement Learning
• Trust Region Policy Optimization with Optimal Transport Discrepancies: Duality and Algorithm for Continuous Actions
• Behavior Regularized Offline Reinforcement Learning
• Policy Optimization as Wasserstein Gradient Flows

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks