MorphAct: Robot Policies Morphing with the Worlds They Act In

Hugging Face Daily Papers Papers
robotics vision-language-action parameter-efficient-finetuning low-rank-adaptation contextual-parameterization imitation-learning

Summary

MorphAct 提出一种上下文参数化框架，通过两个耦合的低秩生成器动态调整机器人策略参数，缓解不同交互上下文中相互冲突的参数更新，在 LIBERO、Meta-World 及真实机器人实验中相比多个 VLA 骨干网络最高提升 16.30%。

Progress toward general intelligence in the physical world requires scaling robot learning across diverse environments and interactions. However, existing robot policies use shared parameters that remain fixed across interaction contexts. When different operations favor competing parameter adjustments, fitting their mixture can leave individual operations insufficiently adapted. Our atomic-task diagnostics reveal competing update directions in both vision-language and action parameters, with source-improving updates increasing loss on other tasks. This motivates MorphAct, a contextual parameterization framework that dynamically adjusts robot policy parameters as interaction contexts change. Two coupled generators produce context-dependent low-rank updates for the vision-language and action modules. Their coupling follows the policy's computational dependency: vision-language updates shape the representations that condition action-side parameter generation. Both generators are trained jointly through the native action objective, without operation labels or expert-adapter targets. MorphAct improves performance across three VLA backbones on LIBERO and Meta-World, with gains of up to 16.30%. In real-world evaluation, it achieves gains of up to 14% over π_{0.5}. Ablations support the benefits of contextual parameter generation and representation coupling.
Original Article
View Cached Full Text

Cached at: 10/19/26, 07:28 AM

Paper page - MorphAct: Robot Policies Morphing with the Worlds They Act In

Source: https://huggingface.co/papers/2610.20019 Published on Oct 14

·

Submitted byhttps://huggingface.co/ziyan2003

liuon Oct 19

Abstract

MorphAct dynamically adapts robot policy parameters via coupled low-rank generators to resolve conflicting updates across interaction contexts, improving vision-language-action performance.

Progress toward general intelligence in the physical world requires scaling robot learning across diverse environments and interactions. However, existing robot policies use shared parameters that remain fixed across interaction contexts. When different operations favor competing parameter adjustments, fitting their mixture can leave individual operations insufficiently adapted. Our atomic-task diagnostics reveal competing update directions in bothvision-languageandaction parameters, with source-improving updates increasing loss on other tasks. This motivates MorphAct, acontextual parameterizationframework that dynamically adjusts robot policy parameters as interaction contexts change. Twocoupled generatorsproduce context-dependentlow-rank updatesfor thevision-languageand action modules. Their coupling follows the policy’s computational dependency:vision-languageupdates shape the representations that condition action-side parameter generation. Both generators are trained jointly through the native action objective, without operation labels or expert-adapter targets. MorphAct improves performance across threeVLA backboneson LIBERO and Meta-World, with gains of up to 16.30%. In real-world evaluation, it achieves gains of up to 14% over π_{0.5}. Ablations support the benefits of contextual parameter generation and representation coupling.

View arXiv pageView PDFProject pageGitHub19Add to collection

Get this paper in your agent:

hf papers read 2610\.20019

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.20019 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2610.20019 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.20019 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

CoDA: Co-Evolution of Data Engines and Action Policies

Hugging Face Daily Papers

CoDA 提出一种仿真数据引擎与机器人动作策略协同演进的框架，通过共享失败与学习进展记忆、视觉语言诊断以及 Text2potential 分阶段势能奖励，将机器人操作任务的分布外成功率从 5.3% 提升至 27.6%，并在 LIBERO-Pro 上取得显著改善。

Co-Evolving Robot Orchestrators and Policies through Deployment

Hugging Face Daily Papers

提出了RoboCOP框架，让机器人编排器(VLM orchestrator)与视觉-语言-动作(VLA)策略在部署过程中协同进化，通过自采集演示数据在验证通过时微调策略，实现自改进飞轮，在仿真和真实世界任务上显著提升了泛化成功率。

Embodied Turing Machines: Stateful Code for Robot Recursive Self-Improvement

Hugging Face Daily Papers

本文提出 Code-Only-As-Policy (COAP) 范式，将机器人本体视为图灵机，用显式代码替代 VLM/VLA 进行决策与状态跟踪，并借助编码智能体实现闭环递归自我改进；在 RoboDojo 的 42 项双臂任务上达到 70.24% 成功率，且测试时无需模型在环。

WAVE-0: Bridging VLA Policies and WAMs with Lightweight Latent Future Prediction

Hugging Face Daily Papers

WAVE-0 提出一个轻量级潜在世界模型，将 VLA 策略与世界动作模型（WAM）结合，使预测的动作后果直接引导动作生成，在 RoboTwin2.0 和 LIBERO 系列基准上取得 91.47%/98.7% 的成功率，并将 π0.5 的真实任务成功率从 38.3% 提升至 73.3%。