mlsanityData to 13 Oct 2026

HorizonFlow: Variable-Length Planning for Offline Goal-Conditioned RL

arXiv 2609.36896 ↗First posted 29 Sep 2026cs.AIcited by 0

On the map: Offline RL and imitation learning from data → Diffusion and flow policies for offline RL

Abstract (arXiv)

Recent advances in generative planning have made trajectory inpainting a promising approach to offline goal-conditioned reinforcement learning. However, these methods typically specify the planning horizon before generating plan content, even though the appropriate horizon depends on the route itself. A horizon that is too short can force infeasible transitions, whereas one that is too long can introduce redundant motion. We introduce HorizonFlow, a hierarchical planner that treats plan length as an output of generation rather than a prescribed input. Its subgoal route planner guides its action-prefix controller through a sequence of latent subgoals. Both components combine insertion-based generation with flow matching to jointly generate continuous plan content and length, using the partially generated plan to guide token insertion. HorizonFlow reuses the resulting length information to select candidates and steer generation toward shorter plans without a separate learned value model. Across Maze2D, Multi2D, and OGBench navigation and visual manipulation benchmarks, HorizonFlow achieves the highest average performance among the compared methods.

Key facts. No citing papers in the map yet (data through 13 Oct 2026). Placed in Diffusion and flow policies for offline RL, one of 13 directions in Offline RL and imitation learning from data. Builds on 23 of its 24 references that are in the map. First posted on arXiv 29 Sep 2026.

Builds on

23 of 24 references.

Closest ancestors · 5

Same direction, newest first

  1. 2026Compositional Transduction with Latent Analogies for Offline Goal-Conditioned Reinforcement Learning20 May
  2. 2025Prior-Guided Diffusion Planning for Offline Reinforcement Learning16 May
  3. 2024OGBench: Benchmarking Offline Goal-Conditioned RL26 Oct
  4. 2024Stitching Sub-Trajectories with Conditional Diffusion Model for Goal-Conditioned Offline RL11 Feb
  5. 2022Is Conditional Generative Modeling all you need for Decision-Making?28 Nov

Other references · 18

  1. 2018Representation Learning with Contrastive Predictive Codingcited by 7,712
  2. 2022Flow Matching for Generative Modelingcited by 5,542
  3. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,130
  4. 2022Planning with Diffusion for Flexible Behavior Synthesiscited by 1,035
  5. 2024Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusioncited by 653
  6. 2024Discrete Flow Matchingcited by 309
  7. 2022Contrastive Learning as Goal-Conditioned Reinforcement Learningcited by 244
  8. 2025A General Framework for Inference-time Scaling and Steering of Diffusion Modelscited by 217
  9. 2023HIQL: Offline Goal-Conditioned RL with Latent States as Actionscited by 138
  10. 2023Optimal Goal-Reaching Reinforcement Learning via Quasimetric Learningcited by 118
  11. 2024Simple Hierarchical Planning with Diffusioncited by 72
  12. 2025Edit Flows: Flow Matching with Edit Operationscited by 53
6 more
  1. 2025Generative Trajectory Stitching through Diffusion Compositioncited by 40
  2. 2023Trans-Dimensional Generative Modeling via Jump Diffusion Modelscited by 34
  3. 2025OneFlow: Concurrent Mixed-Modal and Interleaved Generation with Edit Flowscited by 17
  4. 2026HDFlow: Hierarchical Diffusion-Flow Planning for Long-horizon Taskscited by 2
  5. 2026Variational Learning for Insertion-based Generationcited by 2
  6. 2025VH-Diffuser: Variable Horizon Diffusion Planner for Time-Aware Goal-Conditioned Trajectory Planningcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2022Hierarchical Planning Through Goal-Conditioned Offline Reinforcement Learning
  2. 2026Diffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning
  3. 2026Shortcut Trajectory Planning for Efficient Offline Reinforcement Learning
  4. 2026Bayesian Flow Networks for Offline Trajectory Planning
  5. 2026PathBridger: Subgoal Bridges for Offline Goal-Conditioned Reinforcement Learning
  6. 2025Offline Reinforcement Learning with Generative Trajectory Policies
  7. 2024Diffusion Meets Options: Hierarchical Generative Skill Composition for Temporally-Extended Tasks
  8. 2026MAGE: Multi-scale Autoregressive Generation for Offline Reinforcement Learning