Heteroscedastic Student-t Regression Matches Diffusion Policies in Robot Learning, Paper Finds
YouJiacheng · x · 2026-10-15
An arXiv paper by Yuchen Zhou, Jiacheng You et al. revisits why diffusion/flow-matching policies beat MSE action regression in robot imitation learning. The authors show real demonstration data has state-dependent residual scales and heavier-than-Gaussian tails, and MSE training allocates excess gradient to high-residual observations. Their heteroscedastic Student-t regression (HT-Policies) closes the gap: single-pass inference, no denoising, reusable flow-policy backbones, and success rates competitive with generative baselines across 4 sim benchmarks and real robots.
More from Embodied
- MimicX refines humanoid tracking with policy-in-the-loop feedback: 25.7% error cut — Shuaijun Liu · 2026-10-15
- Gracia's new world model targets Real2Sim2Real and Physical AI — jonstephens85 · 2026-10-15
- Niantic Spatial built VPS on Pokémon Go imagery; a dev already prototyped it solo — jonstephens85 · 2026-10-15
- RoboParty launches RP1, a fully open-source full-stack humanoid robotics platform — GuanyaShi · 2026-10-15
- Survey: Towards Agentic Robots organizes the field into 17 categories across four domains — Chonghan Qin · 2026-10-15
- HumanoidToolBench benchmarks humanoid tool use from selection to mobile execution on real Unitree G1 — snupilab · 2026-10-15