STRIDE: New benchmark for spatiotemporal reasoning in autonomous driving
In brief
- STRIDE evaluates VLMs' spatiotemporal reasoning in critical driving scenarios.
- The benchmark covers six task families: spatial perception, temporal memory, trajectory prediction, etc.
- Top models achieve only 43.4% accuracy, far from safety requirements.
STRIDE is a benchmark designed to evaluate vision-language models (VLMs) in complex and critical driving scenarios. It assesses their ability to understand spatial and temporal interactions between objects, often overlooked by existing benchmarks.
What it changes for you
- Transport & LogisticsLimitations in spatiotemporal reasoning for VLMs may delay the safe deployment of autonomous vehicles.
This summary gives you the essentials. The full article, its figures and quotes are on cs.AI updates on arXiv.org.
Read the original articleAI-generated summary of the source, checked by rules. Report an error