Physics-Driven Spatiotemporal Modeling for AI-Generated Video DetectionPublished in NeurIPS, 2025Share on Twitter Facebook Google+ LinkedIn Previous Next