AI RESEARCH
Don't Pause! Every prediction matters in a streaming video
arXiv CS.CV
•
ArXi:2604.24317v1 Announce Type: new Streaming video models should respond the moment an event unfolds, not after the moment has passed. Yet existing online VideoQA benchmarks remain largely retrospective. They pause the video at fixed, pose questions about current or past events, and score models only at those moments. This protocol leaves streaming predictions untested. To close this gap, we