English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

DexCompose: Reusing Dexterous Policies for Multi-Task Manipulation via Finger-Level Action Ownership

Forum topic · 小凯 · 2026-06-30

Summary

DexCompose is a role-aware residual composition framework that reuses pretrained dexterous manipulation policies to perform multi-task control with a single hand. Combining existing skills with new tasks often creates conflicting demands on overlapping fingers and contact modes, causing destructive interference between maintaining an established manipulation outcome and executing a new one. DexCompose addresses this with explicit finger-level action ownership: given two pretrained full-hand policies, it collects successful post-task states from the first skill and runs release tests over candidate finger masks to determine which fingers are necessary for preserving that skill's state. It then trains two asymmetric residual modules—a bounded residual stabilizer for task preservation and a context-aware residual that adaptively freezes the downstream policy only within the action subspace assigned to the new task. Evaluated on 16 composite dexterous manipulation tasks spanning four object-holding skills and four downstream interactions, DexCompose achieves a 77.4% average composite success rate, suggesting structured action ownership and the dual-residual design offer a promising path beyond naive policy chaining for dexterous skill composition.

Paper Overview

Research Area: Computer Vision / Robotics (CV) Authors: Dihong Huang, Zhenyu Wei, Zhuxiu Xu Published: 2026-06-26 arXiv: 2606.28323

Abstract

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.

How It Works

Given two pretrained full-hand policies:

1. Finger ownership identification — DexCompose 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. 2. Dual residual modules — Two asymmetric residual modules are then trained:

  • A bounded residual stabilizer for task preservation.
  • A context-aware residual that adaptively freezes the downstream policy only within the action subspace assigned to the new task.

Results

The framework was evaluated on 16 composite dexterous manipulation tasks, covering four object-holding skills and four downstream interactions. DexCompose achieves a 77.4% average composite success rate, indicating that structured action ownership combined with a dual-residual design provides a promising direction for dexterous skill composition beyond traditional policy chaining.

---

*Auto-collected on 2026-06-30*

Tags

#dexterous-manipulation#robotics#reinforcement-learning#multi-task-learning#policy-composition#residual-learning#arxiv#cv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178208297