Open Robot Learning Book Adds Planning Chapter, Linking CEM to LLM Test-Time Scaling
GlenBerseth · x · 2026-11-13
Glen Berseth published a new chapter (Lecture 03: Planning & Model-Based RL) of his open Robot Learning book.
- Why planning with a world model is computationally hard, and what structure makes it tractable
- Open-loop vs. closed-loop control under stochastic dynamics
- Sampling-based planners: random shooting, cross-entropy method (CEM), CMA-ES
- Extends the framework to LLMs: best-of-n sampling is random shooting, and recursive self-aggregation lets the model refine its answer distribution, mirroring CEM
The chapter includes videos, slides, a PDF, assignments, and required readings such as AlphaGo Zero, BADGR, and text-guided video policy generation.
Related event: Open Robot Learning Textbook Adds New Planning Chapter(2 posts)→
More from Embodied
- Autonomous dozer factory goes live in Austin, Texas — brezina · 2026-11-13
- Perplexity hiring a Physical Maker to connect robots and AI for fabricating artifacts — inductionheads · 2026-11-13
- Video shows an autonomous parking robot handling the full parking job on its own — DavidWells · 2026-11-13
- BoSS benchmark released: 180K+ stellarators, 3TB of simulation data for generative models — ZhongingAlong · 2026-11-13
- Fly brain hooked up to a synth learns to sing Bad Apple — Bedrovelsen · 2026-11-13
- CoRL takeaways: SuperMimic steals the show, dexterity and tactile sensing improve, human data still a bottleneck — KyleMorgenstein · 2026-11-13