mlsanityData to 14 Oct 2026

Counterfactual Transport Flows for Offline Conservative Trajectory Refinement

arXiv 2606.09115 ↗First posted 8 Jun 2026cs.LGcited by 0

On the map: Offline RL and imitation learning from data → Offline RL algorithms

also in 1 more
Abstract (arXiv)

Offline reinforcement learning (RL) offers a path to policy improvement from logged data alone, using historical returns or other measurable outcomes as world feedback. A key difficulty is improving observed behavior without extrapolating beyond what the offline data supports. We propose counterfactual transport flows, a source-conditioned trajectory refinement framework for offline decision-making guided by world feedback. Given a low-feedback candidate trajectory, we construct local preference pairs from offline data by retrieving nearby trajectories in latent trajectory space with higher task-specific feedback, and use them as weak supervision for conservative refinement. The framework learns instance-specific refinement directions: at inference time, a refinement strength parameter controls how far the candidate trajectory is transported, enabling a trade-off between preserving the original behavior and applying stronger improvement. Experiments on D4RL benchmarks, including AntMaze and MuJoCo tasks, show that our method improves behavior from historical returns as world feedback, while providing interpretable trajectory-level refinement paths.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Offline RL algorithms, one of 13 directions in Offline RL and imitation learning from data. Builds on 18 of its 24 references that are in the map. First posted on arXiv 8 Jun 2026.

Builds on

18 of 24 references.

Closest ancestors · 5

Same direction, newest first

  1. 2021Offline Reinforcement Learning with Implicit Q-Learning12 Oct
  2. 2021A Minimalist Approach to Offline Reinforcement Learning12 Jun
  3. 2020Conservative Q-Learning for Offline Reinforcement Learning8 Jun
  4. 2020Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems4 May
  5. 2020D4RL: Datasets for Deep Data-Driven Reinforcement Learning15 Apr

Other references · 13

  1. 2022Flow Matching for Generative Modelingcited by 5,702
  2. 2022Planning with Diffusion for Flexible Behavior Synthesiscited by 1,061
  3. 2022Building Normalizing Flows with Stochastic Interpolantscited by 995
  4. 2023Improving and generalizing flow-based generative models with minibatch optimal transportcited by 879
  5. 2022Diffusion Policies as an Expressive Policy Class for Offline Reinforcement Learningcited by 562
  6. 2021Reinforcement learning based recommender systems: A surveycited by 188
  7. 2025Flow Q-Learningcited by 184
  8. 2020Sample-Efficient Reinforcement Learning via Counterfactual-Based Data Augmentationcited by 59
  9. 2025Critique of World Modelcited by 25
  10. 2024A Review of Reinforcement Learning in Financial Applicationscited by 13
  11. 2024Preference Alignment with Flow Matchingcited by 11
  12. 2025LeapFactual: Reliable Visual Counterfactual Explanation Using Conditional Flow Matchingcited by 5
1 more
  1. 2025Offline Reinforcement Learning with Generative Trajectory Policiescited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026When Does Trajectory-Level Supervision Permit Efficient Offline Reinforcement Learning?
  2. 2026TOUR: A Trajectory-Level Unlearning Benchmark for Offline Reinforcement Learning
  3. 2026Diffusion Policy Improvement with Proposal-Conditioned Refinement Flows
  4. 2026From Suboptimal to Optimal: One-Step Latent Flow Policies for Offline Reinforcement Learning
  5. 2023Harnessing Mixed Offline Reinforcement Learning Datasets via Trajectory Weighting
  6. 2026Fisher Decorator: Refining Flow Policy via a Local Transport Map
  7. 2024Rethinking Optimal Transport in Offline Reinforcement Learning
  8. 2023Beyond Uniform Sampling: Offline Reinforcement Learning with Imbalanced Datasets