analiza aiqbench
HALO-WA addresses critical failure modes in world-action models by integrating hybrid-attention mechanisms with online reinforcement learning to correct real-world deployment errors. This approach significantly boosts success rates in precision robotic manipulation tasks, demonstrating that lightweight adapters can effectively refine long-horizon action chunks without extensive retraining. The method bridges the gap between simulated generation and physical execution, offering a practical pathway for deploying robust AI agents in unstructured environments.
HALO-WA: Hybrid-Attention Latent-Guided Online Reinforcement Learning for World-Action Models
arXiv:2607.04265v2 Announce Type: replace-cross Abstract: World-action (WA) models can generate long-horizon action chunks for general-purpose robotic manipulation, but they remain vulnerable to calibration, perception, and contact-dynamics errors in real-world precision tasks, often failing in the final few millimeters of alignment or insertion. We propose HALO-WA, a hybrid-attention latent-guided online reinforcement learning (RL) framework for WA models, which leverages latent features and action priors from the WA generation process through a lightweight actor-critic adapter to enable fast online adaptation to real deployment errors. HALO-WA introduces a hybrid-attention structure that preserves the temporal consistency of action chunks while reading task-relevant information from WA latents conditioned on visual context and end-stage correction requirements, thereby producing refined action chunks. We validate HALO-WA on four real-world precision manipulation tasks, where it improves the average success rate from 26.4\% for WA-base to 87.1\%, outperforming the strongest baseline by 19.2 percentage points while requiring only 45--75 minutes of online training per task. To facilitate reproducibility, we further conduct supplementary simulation experiments in RoboTwin and release the code at https://github.com/YeanRoot/HALO-WA.
Abstract:World-action (WA) models can generate long-horizon action chunks for general-purpose robotic manipulation, but they remain vulnerable to calibration, perception, and contact-dynamics errors in real-world precision tasks, often failing in the final few millimeters of alignment or insertion. We propose HALO-WA, a hybrid-attention latent-guided online reinforcement learning (RL) framework for WA models, which leverages latent features and action priors from the WA generation process through a lightweight actor-critic adapter to enable fast online adaptation to real deployment errors. HALO-WA introduces a hybrid-attention structure that preserves the temporal consistency of action chunks while reading task-relevant information from WA latents conditioned on visual context and end-stage correction requirements, thereby producing refined action chunks. We validate HALO-WA on four real-world precision manipulation tasks, where it improves the average success rate from 26.4\% for WA-base to 87.1\%, outperforming the strongest baseline by 19.2 percentage points while requiring only 45--75 minutes of online training per task. To facilitate reproducibility, we further conduct supplementary simulation experiments in RoboTwin and release the code at this https URL.
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2607.04265 [cs.RO]
(or arXiv:2607.04265v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2607.04265 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 263 cuvinte
Subiecte conexe
PEARS: Physical-Prior-Guided Efficient Adaptation via Failure Reasoning and Diffusion Steering for Tactile Manipulation
arXiv · 7 oct. 2026
Digital States 2026: AI Moves From Experiment to Infrastructure
govtech.com · 24 sept. 2026
[AINews] OpenAI reports Navier-Stokes singularity find in 88 hours using Astra-next, roughly 10,000 agents and 130B tokens (>$40M), a contender for second ever Millennium Prize awarded
Latent Space · 9 sept. 2026
Microsoft’s AI chief says superintelligence is near, but won’t take your job
The Verge · 8 iun. 2026