Research map / Video / Video-language understanding
Learning to Watch: Active Video Anomaly Understanding via Interleaved Policy Optimization
Mengjingcheng Mo, Jiaxu Leng, Xinbo Gao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add active closed-loop interaction instead of uniform-sampling one-shot pipeline (ablation) | Anom-π · XD-Violence | AP (%): 65.99 -> 72.29 | Improved |
| add active interaction instead of random-sampling one-shot pipeline (ablation) | Anom-π · XD-Violence | AP (%): 52.54 -> 72.29 | Improved |
| train with iDPO rather than vanilla DPO without curiosity (AEI) (ablation) | Anom-π · XD-Violence | AP (%): 66.57 -> 72.29 | Improved |
| train with iDPO rather than positive-only SFT (ablation) | Anom-π · XD-Violence | AP (%): 56.86 -> 72.29 | Improved |
| add iDPO policy training over zero-shot model (ablation) | Anom-π · XD-Violence | AP (%): 34.92 -> 72.29 | Improved |
| add Think action (ablation) | Anom-π · XD-Violence | AP (%): 65.54 -> 72.29 | Improved |
| add Sample action (ablation) | Anom-π · XD-Violence | AP (%): 67.83 -> 72.29 | Improved |
| add Backtrack action (ablation) | Anom-π · XD-Violence | AP (%): 70.49 -> 72.29 | Improved |
| add Expand action (ablation) | Anom-π · XD-Violence | AP (%): 71.48 -> 72.29 | Improved |
| score trajectories with interaction cost, not final prediction only (ablation) | iDPO · XD-Violence | AP (%): -5.42 AP when scoring by final prediction only; -2.25 when removing interaction cost | Improved |
| set interaction-cost weight eta to 0.5 instead of 0 (no cost) (ablation) | iDPO · XD-Violence | AP (%): 70.04 -> 72.29 | Improved |
| increase eta to 2.0 from 0.5 (ablation) | iDPO · XD-Violence | AP (%): 72.29 -> 69.90 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- VADTree — VADTree: Explainable Training-Free Video Anomaly Detection via Hierarchical Granularity-Aware Tree (NeurIPS 2025) · better on 1 of 2 shared results
- Holmes-VAU — Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity (CVPR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Unified Reasoning Framework for Holistic Zero-Shot Video Anomaly Analysis (NeurIPS 2025 · similarity 0.87)
- Vad-R1: Towards Video Anomaly Reasoning via Perception-to-Cognition Chain-of-Thought (NeurIPS 2025 · similarity 0.87)
- VERA: Explainable Video Anomaly Detection via Verbalized Learning of Vision-Language Models (CVPR 2025 · similarity 0.86)
- Find, Fix, Reason: Context Repair for Video Reasoning (ICML 2026 · similarity 0.86)
- Holmes-VAU: Towards Long-term Video Anomaly Understanding at Any Granularity (CVPR 2025 · similarity 0.86)
- ReAgent-V: A Reward-Driven Multi-Agent Framework for Video Understanding (NeurIPS 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
