EgoLAP paper adds motion-level language reasoning to VLA pretraining, lifting real-world robot task progress to 80.1%
Majumdar_Ani · x · 2026-10-12
A Princeton team (Lihan Zha, Anirudha Majumdar et al.) introduces EgoLAP, a VLA pretraining framework built on the insight that behavior cloning supervises what actions to take but leaves the why buried in trajectories. EgoLAP expresses motion intent as structured, temporally abstracted language actions paired with motion-level reasoning grounded in scene geometry, physics, and object affordances, jointly learning from human egocentric and robot data via a shared language-action chain-of-thought. In real-world and sim experiments it transfers human experience to robot control better than alternative action representations, reaching 80.1% mean real-world task progress — a 2.3x gain — and outperforming composite subtask/object-box/visual-trace reasoning.
More from Embodied
- Cybercab races Waymo on Austin streets as two robotaxis compete in public — Scobleizer · 2026-10-12
- Building a ModRetro game with his daughter and Codex: 5 concrete design fixes — petergyang · 2026-10-12
- Bindu Reddy: superintelligent agent swarms simulating billions of runs will crack mass-market robots — bindureddy · 2026-10-12
- Vernata: ETH Zurich's multi-teacher distillation for self-supervised LiDAR representations, IROS 2026 — rsasaki0109 · 2026-10-12
- Zero real-world data: Success-Guided Sampling unlocks zero-shot dexterous robot manipulation via sim2real RL — Scobleizer · 2026-10-12
- Linux boots on iPhone 11, courtesy of parts hacker — nullagent · 2026-10-12