arXiv:2606. 00702v1 Announce Type: cross Abstract: We propose to turn generalist multi-embodiment value functions into reusable models for robot design.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.
Recent progress in large-scale imitation learning for robot manipulation has been driven by leveraging datasets across a wide range of robot embodiments. However, achieving significant cross-embodiment transfer is often still challenging.
CrossSafe proposes embodiment-conditioned safety filtering that uses a Hamilton‑Jacobi reachability value function shared across robots while conditioning on each robot’s morphology and kinematics via a morphology‑aware latent representation. The method performs reachability analysis directly in latent space, enabling a single policy trained on multiple bimanual robot embodiments and manipulation tasks to generalize zero‑shot to a held‑out embodiment and reduce collision rates. Experiments on five embodiments and five tasks demonstrate that training with more embodiments improves generalization.
The paper discusses how robotic embodiment—sensing, kinematics, dynamics, geometry, actuation, and control—varies across robots and over time, and argues that general embodied intelligence must learn across these differences. It critiques current methods that engineer correspondences for short‑term gains, proposing instead that learning should discover representations that enable transfer across a broader range of embodiments as experience accumulates. The authors advocate for embodiment diversity as a scaling axis, broad learned priors as a complementary ingredient, and evaluations that better characterize embodiment gaps and transfer performance, linking practical cross‑embodiment learning to the scientific pursuit of physical intelligence that adapts with its embodiments.
arXiv:2607. 27549v1 Announce Type: cross Abstract: Recent progress in large-scale imitation learning for robot manipulation has been driven by leveraging datasets across a wide range of robot embodiments.
arXiv:2506. 08630v3 Announce Type: replace Abstract: A universal controller for any robot morphology would greatly improve computational and data efficiency.
The paper discusses the "embodiment gap" in robot foundation models, highlighting that while models can generalize across tasks, additional work is often needed to deploy them on specific robot bodies. It surveys what components can be reused across different robot embodiments and what must be implemented anew, mapping existing methods along axes of shared structure and adaptation stage. The authors propose a reporting framework to better assess adaptation efforts and identify remaining challenges for cross-embodiment learning.