Haifa University paper: online instruction-conditioned mistake detection in video, joint model hits 84.1% F1
UoHaifa · hf · 2026-11-21
A University of Haifa team studies online, instruction-conditioned mistake detection in egocentric videos: using only observed video, task instructions and prior predictions, the model identifies completed action attempts, localizes them in time, and classifies execution as correct, modified, slip, correction or addition; omitted steps are decided after the video ends.
- They repartition EgoPER to include erroneous videos in all splits for supervised learning
- A modular pipeline (separate step identification, temporal localization, execution verification) is compared with a joint single-step model: step-segmentation [email protected] is similar (93.6% vs 92.1%), but the joint model achieves higher binary mistake-detection [email protected] (84.1% vs 80.7%), with the largest gain on slips (56.4% vs 37.5%)
- Component analysis suggests inter-stage error propagation drives the modular approach's lower score; both designs outperform released AEM and ESTANNet checkpoints on videos containing mistakes
More from Research
- Anthropic researcher claims Riemann Hypothesis proof with Claude, backed by 17.4M-line Lean formalization — Kyrannio · 2026-11-21
- Anthropic's Levent Alpöge claims a Riemann Hypothesis proof discovered working with Claude — Kyrannio · 2026-11-21
- Stanford AI100 releases 2026 report on the scaling era and mainstream generative AI — stanfordnlp · 2026-11-21
- New Multi-Agent Coordination Bench finds frontier agents coordinate poorly; only Opus gains — AlexGDimakis · 2026-11-21
- Riemann Hypothesis Result's Dramatically Written Intro Draws Attention — lishali88 · 2026-11-21
- Stanford CS329Z Lecture 9 Released: Optimizing AI Agents via Prompts vs. Weights — stanfordnlp · 2026-11-21