SEU's LeaP: a learnable source prior boosts robot manipulation success by 25.5 points
机器之心 · wechat · 2026-10-03
Wei Xiucai's team at Southeast University proposes LeaP (Learnable source Prior), accepted at CoRL 2026. Instead of starting generative robot policies (diffusion/flow matching) from a standard Gaussian, a lightweight proprioception-driven MLP head (0.21M params) jointly predicts the mean and state-adaptive variance of a diagonal Gaussian source distribution for action generation. On RoboTwin's 15 bimanual tasks it reaches 81.6% average success, +25.5 points over the standard-Gaussian baseline, and 80.0% on three real Franka tasks. Ablations show proprioception-only priors beat visual variants; learning mean plus adaptive variance (85.3%) clearly beats mean-only (78.0%) or fixed noise (62.7%); explicit likelihood and contrastive supervision add complementary gains. The prior transfers across generators (flow matching, diffusion bridge), and code is open-sourced.
More from Embodied
- ETH's Anymal robot completes 10km autonomous alpine hike with VLM-guided ROAM system — ChongZzZhang · 2026-11-19
- iFixit tears down Apple HomeView: lots of glue and 'endless frustration' — Resident-Fall4351 · 2026-11-19
- Removing human video drops robot OOD success from 60% to 7%: generalization comes from human video — gui_penedo · 2026-11-19
- PSYONIC's prosthetic Ability Hand now used by 90+ robotics orgs, raising Series A — DhowCIO · 2026-11-19
- Simulator misspecification breeds overconfidence: sim-to-real study on SPEED+ — Nikolay Dolgov · 2026-11-19
- Doppel world model renders the robot body and beats action-only baselines across 8 embodiments — mangahomanga · 2026-11-19