LIFT: 20-30 Force Traces Teach Pretrained VLA Models Reactive Force Control
机器之心 · wechat · 2026-09-27
A Shanghai Jiao Tong University team (Lu Cewu, Wen Chuan) with QunChu proposed LIFT (Late Reactive Injection of Force for VLA Post-Training), accepted at CoRL 2026 with high scores and open-sourced.
LIFT teaches pretrained VLA models force perception purely via online post-training—no force-data pretraining needed. On towel folding, book insertion, and Hananoi tasks it significantly outperforms vision-only post-training (e.g., 26.7→56.7) with only 20-30 online force traces. It preserves pretrained knowledge via action-expert weight copying and shifted causal attention masks, raises control-loop frequency from 1Hz to 10Hz, and mixes in force-free data to avoid overfitting. The authors argue the paradigm applies to any MoT-based pretrained model.
More from Embodied
- Copy BOM Is Fine, Copying Product Definition Is the Real Worry in Humanoid Robotics — yongqianme · 2026-09-27
- DeepMind and HHMI open-source flybody: a fruit fly simulated joint by joint in MuJoCo — nikola_mr64990 · 2026-09-27
- Dutch nonprofit unveils X17 autonomous litter-picking robot with gripper arm, arriving fall 2026 — MarvinTBaumann · 2026-09-27
- Europe is automating street trash cleanup with robots — MarvinTBaumann · 2026-09-27
- PEARL team heads to IROS 2026 with orals on robot RL and a generalist-vs-specialist debate — GeorgiaChal · 2026-09-27
- Gen-HumanEgo: a 1TB+ egocentric robotics dataset with hand tracking trends on Hugging Face — genrobot2025 · 2026-09-27