analiza aiqbench
PoLAR introduces a structured latent action space using hyperbolic geometry to explicitly encode the extent and mode of robot movements, addressing a key limitation in current visual reconstruction-based methods. By aligning radial dimensions with transition magnitude and directional vectors with movement type, this approach significantly enhances downstream policy performance across both simulation and real-world robotic experiments. The method demonstrates that geometric structuring of latent spaces effectively facilitates the transfer of visual pretraining capabilities to complex robot control tasks.
PoLAR: Factorizing Extent and Mode in Latent Actions for Robot Policy Learning
arXiv:2606.21139v2 Announce Type: replace-cross Abstract: Robot actions vary in both extent and mode: how much change they produce and what kind of change they produce. A latent action space that reflects this structure may make it easier for downstream policies to map latent actions to robot actions. However, existing latent action methods typically rely on visual reconstruction from observation pairs without explicitly encouraging such structure. We introduce Polar Latent Actions with Radial structure (PoLAR), which imposes a radial-direction structure on latent actions, encouraging radius to encode transition extent and direction to retain transition mode. PoLAR uses temporal offsets between observations as a weak proxy for transition extent, encouraging transitions with larger offsets to occupy larger radii. We instantiate this structure in hyperbolic space, whose expanding volume with radius offers a natural fit for more diverse transition modes at larger extents. Across in-task and large-scale pretraining settings, PoLAR improves downstream policy performance in simulation and real-world robot experiments, outperforming latent action baselines and strong pretrained VLAs. These results suggest that structuring latent actions by extent and mode helps transfer visual pretraining to robot policy learning.
Abstract:Robot actions vary in both extent and mode: how much change they produce and what kind of change they produce. A latent action space that reflects this structure may make it easier for downstream policies to map latent actions to robot actions. However, existing latent action methods typically rely on visual reconstruction from observation pairs without explicitly encouraging such structure. We introduce Polar Latent Actions with Radial structure (PoLAR), which imposes a radial-direction structure on latent actions, encouraging radius to encode transition extent and direction to retain transition mode. PoLAR uses temporal offsets between observations as a weak proxy for transition extent, encouraging transitions with larger offsets to occupy larger radii. We instantiate this structure in hyperbolic space, whose expanding volume with radius offers a natural fit for more diverse transition modes at larger extents. Across in-task and large-scale pretraining settings, PoLAR improves downstream policy performance in simulation and real-world robot experiments, outperforming latent action baselines and strong pretrained VLAs. These results suggest that structuring latent actions by extent and mode helps transfer visual pretraining to robot policy learning.
Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2606.21139 [cs.RO]
(or arXiv:2606.21139v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.21139 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 · 276 cuvinte
Subiecte conexe
Links #5: 2026/07
LessWrong · 31 iul. 2026
Learning to Assist: Collaborative VLAs for Implicit Human-Robot Collaboration
arXiv · 12 iun. 2026
[AINews] OpenAI to reach AGI bar by end-2026
Latent Space · 28 aug. 2026
Learning to Assist: Collaborative VLAs for Implicit Human-Robot Collaboration
arXiv · 14 oct. 2026