PINT: A Unified Model with Shared Predictive Intent for Robot Manipulation
PINT: A Unified Model with Shared Predictive Intent for Robot Manipulation
PINT is a unified model that coordinates semantic reasoning, world modeling, and action generation through a shared chunk-level Predictive Intent. It inserts learnable Intent Tokens into the VLM sequence to condition both the VGM and an Action Expert (AE). PINT achieves strong in-distribution performance and significantly improves out-of-distribution generalization, outperforming existing models in tasks like RoboTwin 2.0-Hard and LIBERO-Plus.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ntent and adapts the VGM and AE. Without embodied pretraining, PINT achieves strong in-distribution (ID) performance and substantially improves out-of-distribution (OOD) generalization, reaching 15.5% on RoboTwin 2.0-Hard compared with 1.8% for Fast-WAM and 3.2% for StarVLA, and 79.5% on LIBERO-Plus compared with 51.5% and 75.0%, respectively. Across five real-world bimanual tasks, PINT reaches 68% ID and 24% OOD suc
Open source S5
cess, compared with 35%/9% for pretrained Motus and 67%/35% for pretrained $\pi_{0.5}$.
Open source S6
Source：arXiv Robotics — research abstracts · arxiv.org