arXiv Robotics — research abstracts· Shiva Sreeram, Erfan Aasi, Tsun-Hsuan Wang, Alaa Maalouf, Daniela Rus·· 22 hours agoEditorial score22
Evaluating Video Models for Temporal Understanding in Driving
Mind the Temporal Gap: Evaluating Video Models for Temporal Understanding in Driving
Summary
Video-language models show unreliable temporal reasoning in driving tasks. Lightweight finetuning and probes outperform prompting methods, suggesting motion cues are present but not effectively utilized.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org