Hydra-0: Action Flow for Generalist World Modeling and Control

Hugging Face Daily Papers Papers

Summary

Hydra-0 introduces action flow as a shared visual interface for generalist world modeling and robot control, achieving significant error reductions and enabling zero-shot composition across diverse embodiments.

We introduce Hydra-0, a generalist world model conditioned on action flow, which represents robot actions as pixel motion. This shared visual interface enables generalist world modeling and control by learning action consequences across embodiments, tasks, environments, and video-generation backbones. Our best configuration achieves 90.4% lower robot-motion error and 60.2% lower object-motion error than our action-conditioned baseline, while supporting zero-shot composition and data-efficient adaptation. On the RoboLab benchmark, Hydra-0 achieves a Pearson correlation of r=0.96 between replayed and reference success rates. Finally, we uncover an emergent inverse mode of this interface: a world action model that predicts compatible robot motion from desired object flow transferred from a human demonstration. A trained action head maps the resulting latent features to executable actions without requiring task-specific expert robot demonstrations. Together, these results demonstrate the potential of action flow as a shared control interface connecting heterogeneous training data, open-loop policy evaluation, and robot control.
Original Article
View Cached Full Text

Cached at: 08/24/26, 04:29 PM

Paper page - Hydra-0: Action Flow for Generalist World Modeling and Control

Source: https://huggingface.co/papers/2608.18077 Authors:

,

,

,

,

,

,

,

,

,

Abstract

Hydra-0 uses action flow as a shared visual interface for generalist world modeling and robot control across diverse embodiments and tasks.

We introduce Hydra-0, a generalistworld modelconditioned onaction flow, which represents robot actions aspixel motion. This shared visual interface enables generalistworld modeling and control by learning action consequences across embodiments, tasks, environments, and video-generation backbones. Our best configuration achieves 90.4% lower robot-motion error and 60.2% lower object-motion error than our action-conditioned baseline, while supportingzero-shot compositionand data-efficient adaptation. On the RoboLab benchmark, Hydra-0 achieves a Pearson correlation of r=0.96 between replayed and reference success rates. Finally, we uncover an emergentinverse modeof this interface: aworld action modelthat predicts compatible robot motion from desired object flow transferred from a human demonstration. A trained action head maps the resulting latent features to executable actions without requiring task-specific expert robot demonstrations. Together, these results demonstrate the potential ofaction flowas a shared control interface connecting heterogeneous training data,open-loop policy evaluation, and robot control.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2608\.18077

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/2608.18077 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.18077 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.18077 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

Masked Visual Actions for Unified World Modeling

Hugging Face Daily Papers

Introduces Masked Visual Actions, a pixel-space control interface that expresses actions as partially revealed trajectories, enabling a single model to act as forward dynamics model, recover robot behavior, and support model-based planning and inverse modeling with only 15 hours of training data.

AHA-WAM:Asynchronous Horizon-Adaptive World-Action Modeling with Observation-Guided Context Routing

Hugging Face Daily Papers

AHA-WAM is an asynchronous world-action model that uses dual Diffusion Transformers to decouple world prediction from action execution, achieving efficient long-horizon planning and real-time control. It achieves state-of-the-art performance on robotic manipulation tasks with up to 92.8% success on RoboTwin and 78.3% on real-world tasks, while reaching 24.17 Hz closed-loop control.