DeCo: Task Decomposition and Skill Composition for Zero-Shot Generalization in Long-Horizon 3D Manipulation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Zixuan, Yin, Junhui, Chen, Yangtao, Huo, Jing, Tian, Pinzhuo, Shi, Jieqi, Hou, Yiwen, Li, Yinchuan, Gao, Yang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914330176389120
author Chen, Zixuan
Yin, Junhui
Chen, Yangtao
Huo, Jing
Tian, Pinzhuo
Shi, Jieqi
Hou, Yiwen
Li, Yinchuan
Gao, Yang
author_facet Chen, Zixuan
Yin, Junhui
Chen, Yangtao
Huo, Jing
Tian, Pinzhuo
Shi, Jieqi
Hou, Yiwen
Li, Yinchuan
Gao, Yang
contents Generalizing language-conditioned multi-task imitation learning (IL) models to novel long-horizon 3D manipulation tasks is challenging. To address this, we propose DeCo (Task Decomposition and Skill Composition), a model-agnostic framework that enhances zero-shot generalization to compositional long-horizon manipulation tasks. DeCo decomposes IL demonstrations into modular atomic tasks based on gripper-object interactions, creating a dataset that enables models to learn reusable skills. At inference, DeCo uses a vision-language model (VLM) to parse high-level instructions, retrieve relevant skills, and dynamically schedule their execution. A spatially-aware skill-chaining module ensures smooth, collision-free transitions between skills. We introduce DeCoBench, a benchmark designed to evaluate compositional generalization in long-horizon manipulation tasks. DeCo improves the success rate of three IL models, RVT-2, 3DDA, and ARP, by 66.67%, 21.53%, and 57.92%, respectively, on 12 novel tasks. In real-world experiments, the DeCo-enhanced model, trained on only 6 atomic tasks, completes 9 novel tasks in zero-shot, with a 53.33% improvement over the baseline model. Project website: https://deco226.github.io.
format Preprint
id arxiv_https___arxiv_org_abs_2505_00527
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DeCo: Task Decomposition and Skill Composition for Zero-Shot Generalization in Long-Horizon 3D Manipulation
Chen, Zixuan
Yin, Junhui
Chen, Yangtao
Huo, Jing
Tian, Pinzhuo
Shi, Jieqi
Hou, Yiwen
Li, Yinchuan
Gao, Yang
Robotics
Generalizing language-conditioned multi-task imitation learning (IL) models to novel long-horizon 3D manipulation tasks is challenging. To address this, we propose DeCo (Task Decomposition and Skill Composition), a model-agnostic framework that enhances zero-shot generalization to compositional long-horizon manipulation tasks. DeCo decomposes IL demonstrations into modular atomic tasks based on gripper-object interactions, creating a dataset that enables models to learn reusable skills. At inference, DeCo uses a vision-language model (VLM) to parse high-level instructions, retrieve relevant skills, and dynamically schedule their execution. A spatially-aware skill-chaining module ensures smooth, collision-free transitions between skills. We introduce DeCoBench, a benchmark designed to evaluate compositional generalization in long-horizon manipulation tasks. DeCo improves the success rate of three IL models, RVT-2, 3DDA, and ARP, by 66.67%, 21.53%, and 57.92%, respectively, on 12 novel tasks. In real-world experiments, the DeCo-enhanced model, trained on only 6 atomic tasks, completes 9 novel tasks in zero-shot, with a 53.33% improvement over the baseline model. Project website: https://deco226.github.io.
title DeCo: Task Decomposition and Skill Composition for Zero-Shot Generalization in Long-Horizon 3D Manipulation
topic Robotics
url https://arxiv.org/abs/2505.00527