X-DiffVLA: X-Embodied Diffusion Action Heads for Vision-Language-Action Models

2026-05-24

AI Summary

This work addresses the limited generalization of vision-language-action (VLA) models across embodied agents and their reliance on embodiment-specific fine-tuning by proposing X-DiffVLA, a unified diffusion-based VLA framework. The approach introduces Embodiment Forcing guidance and Morphological Tree Diffusion to implicitly model embodiment-specific structures without explicit supervision, while enhancing behavioral coherence across heterogeneous end-effectors. Leveraging classifier-free guidance and a shared cross-embodiment action head, X-DiffVLA effectively captures both the diversity and latent consistency inherent in cross-embodiment data. Experimental results demonstrate significant improvements, with task success rates increasing by 15.3% on RoboCasa and 12.5% on Isaac Gym, and real-world evaluations confirm its robustness and scalable capacity for cross-embodiment policy learning.

Abstract

Learning universal policies from cross-embodied data remains a fundamental challenge in robotics. Although Vision-Language-Action (VLA) models are pre-trained on large and diverse datasets, they typically rely on embodiment-specific fine-tuning to achieve strong performance in downstream tasks. This requirement severely limits their generalization capability and restricts knowledge transfer across embodiments performing similar tasks. To overcome these limitations, we focus on cross-embodied settings with shared robotic bases and heterogeneous end-effectors, and propose X-DiffVLA, a diffusion-based VLA model featuring a unified cross-embodied action head. X-DiffVLA can leverage the generative strengths of diffusion models to capture both the diversity and latent correlations in cross-embodied datasets. Specifically, we introduce Embodiment Forcing, a classifier-free guidance technique to implicitly steer action generation toward embodiment-specific functional components, capturing fine-grained structural nuances without explicit supervision. In addition, a Morphological Tree Diffusion approach is designed to strengthen behavioral correlations across diverse end-effectors, maximizing the transferability of heterogeneous demonstrations. Experimental results across RoboCasa and Isaac Gym, covering different embodiments from grippers to dexterous hands, show that X-DiffVLA achieves state-of-the-art performance, with improvements of 15.3% and 12.5%, respectively. Real-world evaluations further validate the robustness of the proposed framework and its effectiveness in scalable cross-embodied policy learning.

Problem

• cross-embodied learning
• universal policy
• vision-language-action models
• generalization
• embodiment transfer

Innovation

• Cross-embodied learning
• Diffusion models
• Vision-Language-Action (VLA)
• Embodiment Forcing
• Morphological Tree Diffusion