HuGo: Berkeley Team's LLM-Designed Humanoid Whole-Body Policy Paper Spotlit at CoRL 2026
m_wulfmeier · x · 2026-11-10
mwulfmeier congratulates collaborator Negar Mehr on a CoRL 2026 early-career spotlight (Berkeley AI). The conference slide shows their joint paper HuGo: "LLMs as Whole-Body Policy Code Designers for Humanoid Loco-Manipulation" (arXiv 2609.32108, authors include Seoyeon Choi et al.).
The approach uses a high-level "Coach Policy Code" module that maps observations to high-level actions driving a frozen low-level whole-body policy, effectively coaching existing humanoid skills onto new tasks — with LLMs acting as designers of the policy code.
More from Embodied
- Scale AI's Muse assistant launches on Apple Watch and WearOS — SiddDevs · 2026-11-10
- MuJoCo Playground 1.0 ships full sim-to-real workflows with SO101 and YAM robots supported out of the box — philfung · 2026-11-10
- Synthetic wrist-camera renders for robot soft-object poking look trippy — andrew_n_carr · 2026-11-10
- Apple paper: generalist AI model drives robot arm zero-shot at 85.6%, crushing specialist VLA policies — rohanpaul_ai · 2026-11-10
- Boston Dynamics puts roller skates on Atlas, RL nails mid-air 360 spin on first try — LinusEkenstam · 2026-11-10
- Musk: Cybertruck Orders Exceed Production, Line Shuts in ~8 Months as Focus Shifts to Cybercab — minchoi · 2026-11-10