Research map / Video / Action recognition
UIL-AQA: Uncertainty-Aware Clip-Level Interpretable Action Quality Assessment
Xu Dong, Xinran Liu, Wanqing Li, Anthony Olufemi Tesimi Adeyemi-Ejeye, Andrew Gilbert
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 attention loss (ablation) | query-based decoder baseline · RG avg | SRCC: 0.628 -> 0.807 | Improved |
| add query positional encoding (ablation) | decoder + attention loss · RG avg | SRCC: 0.807 -> 0.810 | Improved |
| add query initialization with high variance (ablation) | decoder + attention loss + query PE · RG avg | SRCC: 0.810 -> 0.842 | Improved |
| add memory positional encoding to query PE (ablation) | UIL-AQA · RG avg | SRCC: 0.824 -> 0.751 | Hurt |
| use memory PE only instead of query PE (ablation) | UIL-AQA · RG avg | SRCC: 0.824 -> 0.778 | Hurt |
| increase query init variance from 0.5 to 5 (ablation) | UIL-AQA · RG avg | SRCC: 0.810 -> 0.820 | Improved |
| increase query init variance from 1 to 3 (ablation) | UIL-AQA · RG avg | SRCC: 0.810 -> 0.811 | No clear effect |
| use UIL-AQA instead of GDLT (ablation) | VST features · RG avg | SRCC: 0.765 -> 0.842 | Improved |
| use UIL-AQA instead of GDLT | VST features · Fis-V avg | SRCC: 0.761 -> 0.788 | Improved |
| use UIL-AQA instead of CoRe-GOAT | VST features · LOGO | SRCC: 0.560 -> 0.780 | Improved |
| use UIL-AQA instead of CoRe-GOAT | I3D features · LOGO | SRCC: 0.494 -> 0.593 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ACTION-NET — ACTION-Net: Multipath Excitation for Action Recognition (CVPR 2021) · 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
- Action Quality Assessment with Temporal Parsing Transformer (ECCV 2022 · similarity 0.92)
- A Decade of Action Quality Assessment: Largest Systematic Survey of Trends, Challenges, and Future Directions (IJCV 2026 · similarity 0.89)
- What and How Well You Performed? A Multitask Learning Approach to Action Quality Assessment (CVPR 2019 · similarity 0.88)
- GAIA: Rethinking Action Quality Assessment for AI-Generated Videos (NeurIPS 2024 · similarity 0.86)
- FineParser: A Fine-grained Spatio-temporal Action Parser for Human-centric Action Quality Assessment (CVPR 2024 · similarity 0.86)
- LucidAction: A Hierarchical and Multi-model Dataset for Comprehensive Action Quality Assessment (NeurIPS 2024 · 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.
