mlsanityData to 7 Oct 2026

DexCompose: Reusing Dexterous Policies for Multi-Task Manipulation with a Single Hand

arXiv 2606.28323 ↗First posted 26 Jun 2026cs.RO cs.AI cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Dexterous manipulation policies can solve individual skills, but composing them to perform multiple tasks with a single hand remains challenging. Adding a new task on top of an existing manipulation skill often imposes conflicting demands on overlapping fingers and contact modes, causing destructive interference between preserving an existing manipulation outcome and executing a new one. We propose DexCompose, a role-aware residual composition framework that reuses pretrained dexterous policies for multi-task manipulation through explicit finger-level action ownership. Given two pretrained full-hand policies, DexCompose first collects successful post-task states from the first skill and performs release tests over candidate finger masks to identify which fingers are necessary for maintaining the established skill state. It then trains two asymmetric residual modules: a bounded residual stabilizer for task preservation, and a context-aware residual that adapts the frozen downstream policy only within the action subspace assigned to the new task. We evaluate the framework on 16 composite dexterous manipulation tasks spanning four object-retention skills and four downstream interactions. DexCompose achieves a 77.4% average composite success rate, demonstrating that structural action ownership with dual residuals offers a promising direction for composing dexterous skills beyond conventional policy chaining.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 38 of its 41 references that are in the map. First posted on arXiv 26 Jun 2026.

Builds on

38 of 41 references.

References · 38

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,730
  3. 2019Solving Rubik's Cube with a Robot Handcited by 699
  4. 2017Learning Complex Dexterous Manipulation with Deep Reinforcement Learning and Demonstrationscited by 685
  5. 2018Learning Dexterous In-Hand Manipulationcited by 643
  6. 2018Residual Reinforcement Learning for Robot Controlcited by 242
  7. 2020Accelerating Reinforcement Learning with Learned Skill Priorscited by 220
  8. 2021DexMV: Imitation Learning for Dexterous Manipulation from Human Videoscited by 194
  9. 2018Residual Policy Learningcited by 146
  10. 2020Parrot: Data-Driven Behavioral Priors for Reinforcement Learningcited by 136
  11. 2022DexGraspNet: A Large-Scale Robotic Dexterous Grasp Dataset for General Objects Based on Simulationcited by 129
  12. 2023UniDexGrasp: Universal Robotic Dexterous Grasping via Learning Diverse Proposal Generation and Goal-Conditioned Policycited by 117
26 more
  1. 2024DexMimicGen: Automated Data Generation for Bimanual Dexterous Manipulation via Imitation Learningcited by 104
  2. 2023UniDexGrasp++: Improving Dexterous Grasping Policy Learning via Geometry-aware Curriculum and Iterative Generalist-Specialist Learningcited by 101
  3. 2022DexVIP: Learning Dexterous Grasping with Human Hand Pose Priors from Videocited by 83
  4. 2024From Imitation to Refinement -- Residual RL for Precise Assemblycited by 80
  5. 2022Learning and Retrieval from Prior Data for Skill-based Imitation Learningcited by 77
  6. 2021Augmenting Reinforcement Learning with Behavior Primitives for Diverse Manipulation Taskscited by 75
  7. 2022Dexterous Imitation Made Easy: A Learning-Based Framework for Efficient Dexterous Manipulationcited by 75
  8. 2023DexArt: Benchmarking Generalizable Dexterous Manipulation with Articulated Objectscited by 64
  9. 2017Data-efficient Deep Reinforcement Learning for Dexterous Manipulationcited by 61
  10. 2025ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learningcited by 56
  11. 2023Sequential Dexterity: Chaining Dexterous Policies for Long-Horizon Manipulationcited by 43
  12. 2024DexGraspNet 2.0: Learning Generative Dexterous Grasping in Large-scale Synthetic Cluttered Scenescited by 41
  13. 2022Iterative Residual Policy: for Goal-Conditioned Dynamic Manipulation of Deformable Objectscited by 36
  14. 2021Residual Reinforcement Learning from Demonstrationscited by 27
  15. 2020Residual Policy Learning for Shared Autonomycited by 25
  16. 2022Residual Skill Policies: Learning an Adaptable Skill-based Action Space for Reinforcement Learning for Roboticscited by 23
  17. 2023Grasp Multiple Objects with One Handcited by 18
  18. 2024DexHandDiff: Interaction-aware Diffusion Planning for Adaptive Dexterous Manipulationcited by 15
  19. 2026UniDex: A Robot Foundation Suite for Universal Dexterous Hand Control from Egocentric Human Videoscited by 12
  20. 2024DexH2R: Task-oriented Dexterous Manipulation from Human to Robotscited by 11
  21. 2026One Hand to Rule Them All: Canonical Representations for Unified Dexterous Manipulationcited by 7
  22. 2021Residual Feedback Learning for Contact-Rich Manipulation Tasks with Uncertaintycited by 6
  23. 2025Sequential Multi-Object Grasping with One Dexterous Handcited by 4
  24. 2025Grasping a Handful: Sequential Multi-Object Dexterous Grasp Generationcited by 3
  25. 2020Proactive Action Visual Residual Reinforcement Learning for Contact-Rich Tasks Using a Torque-Controlled Robotcited by 2
  26. 2026CoorDex: Coordinating Body and Hand Priors for Continuous Dexterous Humanoid Loco-Manipulationcited by 2

Built on by

Not cited yet.