Research map / Video / Action recognition
Procedural Mistake Detection via Action Effect Modeling
Wenliang Guo, Yujiang Pu, Yu Kong
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 action effect modeling with supervised effect tokens (ablation) | AEM · EgoPER | AUC: 67.6 -> 73.8 | Improved |
| add contrastive alignment in state and relation spaces (ablation) | AEM · EgoPER | AUC: 66.8 -> 73.8 | Improved |
| add contrastive alignment in relation space only (ablation) | AEM · EgoPER | AUC: 66.8 -> 72.6 | Improved |
| add contrastive alignment in state space only (ablation) | AEM · EgoPER | AUC: 66.8 -> 69.9 | Improved |
| use proposed effect-frame sampling instead of last frame (ablation) | AEM · EgoPER | AUC: 70.6 -> 73.8 | Improved |
| add dynamic cross-level fusion (ablation) | AEM · EgoPER | AUC: 71.8 -> 73.8 | Improved |
| add dynamic cross-level fusion (ablation) | AEM · EgoPER | EDA: 63.4 -> 66.7 | Improved |
| use relation supervision instead of state supervision (ablation) | AEM · EgoPER | AUC: 68.4 -> 69.4 | Improved |
| use visual instead of textual supervision features (ablation) | AEM · EgoPER | AUC: 69.9 -> 71.7 | Improved |
| use learnable prompts instead of fixed prompts (ablation) | AEM · EgoPER | AUC: 72.8 -> 73.8 | Improved |
| replace GPT-4o with Qwen3-VL for scene graphs (ablation) | AEM · EgoPER | AUC: 73.8 -> 73.3 | No clear effect |
| replace GPT-4o with GPT-5 for scene graphs (ablation) | AEM · EgoPER | AUC: 73.8 -> 74.0 | Improved |
Closest papers by meaning
- Modeling Multiple Normal Action Representations for Error Detection in Procedural Tasks (CVPR 2025 · similarity 0.89)
- Error Detection in Egocentric Procedural Task Videos (CVPR 2024 · similarity 0.88)
- Differentiable Task Graph Learning: Procedural Activity Representation and Online Mistake Detection from Egocentric Videos (NeurIPS 2024 · similarity 0.87)
- PREGO: Online Mistake Detection in PRocedural EGOcentric Videos (CVPR 2024 · similarity 0.87)
- Gazing Into Missteps: Leveraging Eye-Gaze for Unsupervised Mistake Detection in Egocentric Videos of Skilled Human Activities (CVPR 2025 · similarity 0.85)
- When Actions Go Off-Task: Detecting and Correcting Misaligned Actions in Computer-Use Agents (ICML 2026 · similarity 0.85)
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.
