EMPIRIC teaches robots missing physics as code, solving all 25 tasks where baselines get 14-16
tomssilver · x · 2026-10-02
EMPIRIC, from Basis Research Institute with Cambridge, Princeton, MIT and others, learns residual world models: it extends a physics engine with code for missing mechanisms (glue curing, water heating, wind), uses Bayesian inference to estimate parameters from a few noisy experiments, and plans with the result. It solves all 25 tasks across five simulated domains where coding-agent baselines solve 14-16, and on a real robot it learns wind force and domino masses from just two gusts. Paper and code are public.
More from Embodied
- RTX Spark laptops and mini desktops rumored Oct 7 launch, $1800-$2900 with 24GB-128GB — Porespellar · 2026-10-03
- Stanford revisits 2018 claim that self-driving was 90% done — the last 10% was everything — StanfordHAI · 2026-10-03
- City Farmers joins OpenAI's first Codex Physical Builds batch with a Raspberry Pi setup — broodsugar · 2026-10-02
- Five cars on one Austin block, four of them self-driving — binarybits · 2026-10-02
- Tesla adds 32 Cybercabs in Texas as its robotaxi rollout finally scales up — binarybits · 2026-10-02
- Agility's Digit runs end-to-end autonomous whole-body manipulation for ~12 hours at IROS — chris_j_paxton · 2026-10-02