Mao Jiayuan presents RAPID and TR-guided adaptive guidance for robot manipulation at IROS 2026
Overview
Mao Jiayuan of UPenn, with MIT and Stanford, presented the RAPID framework at IROS 2026.
RAPID learns non-grasping manipulation from a single human demonstration. Mao also introduced the TR metric, which measures whether a video-action model's action head actually uses the future frames it predicts.
Mao says adaptive guidance based on TR raised dual-arm real-robot generalization on complex composite tasks to 83%. This figure comes from her presentation as reported by Leiphone, and the report does not include the baseline or the test setup.
Written by AI from the articles below · updated Oct 14, 2:20 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- LeiphoneNewsUPenn's Mao Jiayuan details RAPID and a TR metric that lifts real-robot generalization to 83%
AIMao Jiayuan of UPenn, MIT and Stanford presented the RAPID framework at IROS 2026, which learns non-grasping manipulation from a single human demonstration. She also introduced the TR metric, which measures whether a video-action model's action head actually uses predicted future frames, and says adaptive guidance based on TR raised dual-arm real-robot generalization on complex composite tasks to 83%.
Heat trend
Not enough continuous observations to show a trend yet.