aiqbench analysis
DexTacWAM advances AI robotics by integrating tactile feedback directly into video diffusion world models, enabling more accurate prediction of contact dynamics during dexterous manipulation. This approach allows pretrained vision models to adapt to touch with minimal data while maintaining high-fidelity visual predictions, marking a significant step toward robust embodied intelligence systems.
DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation
arXiv:2609.24976v1 Announce Type: cross Abstract: Dexterous manipulation depends on contact dynamics that are often only partially observable from vision. Recent World-Action Models (WAMs) couple predictive video world modeling with action generation, but remain largely vision-centric and therefore cannot directly model these contact dynamics. We present DexTacWAM, a visuo-tactile WAM that encodes each fingertip independently, aggregates the resulting features through a finger- and pose-aware tactile compressor, and injects the tactile latent into a video diffusion world model for joint visuo-tactile world modeling. Across six contact-rich dexterous manipulation tasks on a 22-DoF bimanual platform, DexTacWAM achieves the highest score on every task, averaging 70.6 versus 38.0 for the strongest baseline. Ablations attribute the gain to modeling contact evolution as part of the predicted world state rather than tactile conditioning alone: removing tactile world modeling reduces the four-task mean from 74.7 to 26.6 while keeping the same tactile features and action expert. After four hours of tactile-encoder adaptation with a frozen pretrained vision VAE, our continual vision-to-touch learning extends the pretrained video model to touch using roughly 100 demonstrations per task without tactile midtraining, while retaining visual prediction quality within 0.5 dB of vision-only counterparts. The compressor retains 89.4% of pre-fusion contact recall while enabling 2.26x faster training and 1.29x faster inference. Together, these results show that pretrained video priors can be extended to distributed multi-finger contact dynamics in a data- and compute-efficient manner.
Abstract:Dexterous manipulation depends on contact dynamics that are often only partially observable from vision. Recent World-Action Models (WAMs) couple predictive video world modeling with action generation, but remain largely vision-centric and therefore cannot directly model these contact dynamics. We present DexTacWAM, a visuo-tactile WAM that encodes each fingertip independently, aggregates the resulting features through a finger- and pose-aware tactile compressor, and injects the tactile latent into a video diffusion world model for joint visuo-tactile world modeling. Across six contact-rich dexterous manipulation tasks on a 22-DoF bimanual platform, DexTacWAM achieves the highest score on every task, averaging 70.6 versus 38.0 for the strongest baseline. Ablations attribute the gain to modeling contact evolution as part of the predicted world state rather than tactile conditioning alone: removing tactile world modeling reduces the four-task mean from 74.7 to 26.6 while keeping the same tactile features and action expert. After four hours of tactile-encoder adaptation with a frozen pretrained vision VAE, our continual vision-to-touch learning extends the pretrained video model to touch using roughly 100 demonstrations per task without tactile midtraining, while retaining visual prediction quality within 0.5 dB of vision-only counterparts. The compressor retains 89.4% of pre-fusion contact recall while enabling 2.26x faster training and 1.29x faster inference. Together, these results show that pretrained video priors can be extended to distributed multi-finger contact dynamics in a data- and compute-efficient manner.
22 pages. Project website: this https URL
Robotics (cs.RO); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
arXiv:2609.24976 [cs.RO]
(or arXiv:2609.24976v1 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2609.24976 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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.
Full text under the publisher's open licence · 335 words
Related coverage
DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination
arXiv · Sep 9, 2026
OmniVTLA: Vision-Tactile-Language-Action Models with Semantic-Aligned Tactile Sensing
arXiv · Jun 16, 2026
ViTacWorld: Scaling Visuo-Tactile World Models for Contact-Rich Robot Manipulation
arXiv · Jul 27, 2026
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026