AskStream benchmark shows models struggle to ask questions proactively in streaming video
CASIA · hf · 2026-10-30
Researchers from CASIA's Institute of Automation, BAAI and partners introduce AskStream, a benchmark for proactive question-asking in streaming video understanding: when information is missing or uncertain, models should decide when and what to ask rather than respond passively.
Moving beyond offline response paradigms toward streaming, proactive interaction, the benchmark reveals that current multimodal models struggle with this decision-making, exposing a key weakness in agent-style video understanding.
More from Research
- Prove2Me formalizes 2,000+ operations research papers in Lean 4 with Claude agents, 64% done — ChrSzegedy · 2026-10-30
- 1-16 Coding Agents Tested: Teams Burn 36% More Tokens to Save 5 Hours — kyang8093 · 2026-10-30
- Noninvasive BCI lets unresponsive ICU patients communicate via brainwaves — ZeroStateReflex · 2026-10-30
- Co-Design Bench: top coding agents fully satisfy only 26.4% of real-user tasks — Qwen · 2026-10-30
- Adobe's EditAgent routes image edits across editors with evidence-based listwise judging — adobe · 2026-10-30
- UniReps 2026 workshop set for December 12 in Paris — ClementineDomi6 · 2026-10-30