HT-Policy Uses Heavy-Tailed Student-t Regression to Close Robot Policy Gap
chris_j_paxton · x · 2026-11-12
Researchers from UPenn, Tsinghua and UCSD revisit the gap between MSE-Policies and Flow-Policies through statistical modeling, analyzing how action-prediction residuals shape policy optimization.
Analysis of human demonstration data reveals state-dependent residual scales and heavier-than-Gaussian tails. They introduce heteroscedastic Student-t action regression (HT-Policies), which learns input-dependent residual scales and reduces the influence of heavy tails. HT-Policies predict action chunks in a single feed-forward pass and match generative policy baselines across four simulation benchmarks and real-robot evaluations, while training and inferring faster.
More from Embodied
- A closer look at Astra's robot gripper design (and a joke at its expense) — _Stocko_ · 2026-11-12
- Robot startup passes first presale milestone with 10+ units claimed — R4nsack · 2026-11-12
- Runway launches Praxis-1, a 27B MoE world action model, with a 73ms Fast variant — ChongZzZhang · 2026-11-12
- Dream-1 on bimanual YAM robot folds a towel on voice command at CoRL — DJiafei · 2026-11-12
- Robot build test: GPT-6 Astra stabilizes two-board structure, Claude Fable 5.2 falls short — qinzytech · 2026-11-12
- Team claims a robot priced like a modern smartphone, presale buyers get the best deal — R4nsack · 2026-11-12