MorphAct: coupled low-rank generators dynamically adapt robot policies across contexts
SJTU · hf · 2026-10-19
Researchers from USTC, SJTU, and NUS propose MorphAct to fix a key limitation of robot policies: shared parameters stay fixed across interaction contexts, and tasks favor competing parameter updates—source-improving updates increase loss on other tasks, diagnostics show conflicts in both vision-language and action parameters.
MorphAct is a contextual parameterization framework using two coupled low-rank generators to dynamically morph policy parameters as interaction contexts change, resolving conflicting updates and improving vision-language-action performance.
More from Embodied
- Asimov open-sources Python SDK and Edge control layer for its humanoid robot — freelerobot · 2026-10-19
- Tesla Roadster Stunts Simulated in MuJoCo with Thruster and Adhesion Actuators — kevin_zakka · 2026-10-19
- BridgeWAM: latent bridge queries decouple video and action experts, 2.47x faster robot inference — Fudan-University · 2026-10-19
- Dev builds a hand-tracking game in hours with Tesla Specs glasses' CLAD — stspanho · 2026-10-19
- Does a robot's jump count as flight? X users pedantically debate the definition — basedjensen · 2026-10-19
- MolmoAct2 Lead Duan Jiafei Joins NUS, Wins Multimillion-Dollar Embodied AI Grant — DJiafei · 2026-10-19