TMLR Survey: The Embodiment Gap in Robot Foundation Models
HirokatuKataoka · x · 2026-08-20
This survey paper published in TMLR introduces the concept of the "embodiment gap"—the work required to adapt reusable foundation models, representations, or data for execution on a robot with a specific body. It maps existing methods on a two-axis diagram showing shared structure types and adaptation stages. The paper examines research directions like sharing semantics, robot data, and learning cross-embodiment correspondences, proposing a reporting framework to evaluate adaptation work beyond just success rates.
More from Embodied
- Kinex: a harness that connects coding agents to any physical robot for scalable learning — daibond_alpha · 2026-10-16
- Gemini is a robotics outlier: dropping it lifts capability-robotics correlation from 0.20 to 0.96 — ycombinator · 2026-10-16
- EPFL team uses RFdiffusion to design de novo metamorphic proteins switching between two folds — ferruz_noelia · 2026-10-16
- Kyma bills itself as the world's most accessible brain-computer interface, opens founding batch preorders — Scobleizer · 2026-10-16
- EvenRealities Accused of Abandoning $1,000 G1 Glasses Owners With Incompatible New App — RileyRalmuto · 2026-10-16
- Amazon's Panos Panay says AI devices will replace the need to use apps — TechCrunch AI · 2026-10-16