Research map / Video / Video-language understanding
Embracing Uncertainty: Decoupling and De-Bias for Robust Temporal Grounding
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use DeNet instead of the strongest listed baseline | Charades-STA temporal grounding · Charades-STA | R@1 IoU=0.7: 35.48 -> 38.52 | Helped |
| Use DeNet instead of the strongest listed baseline | Charades-STA temporal grounding · Charades-STA | R@5 IoU=0.7: 60.05 -> 66.83 | Helped |
| Use DeNet instead of the strongest listed baseline | ActivityNet Captions temporal grounding · ActivityNet Captions | R@1 IoU=0.5: 43.95 -> 43.79 | No clear effect |
| Use DeNet instead of the strongest listed baseline | ActivityNet Captions · ActivityNet Captions | R@5 IoU=0.3: 82.64 -> 86.02 | Helped |
| Use DeNet instead of the strongest listed baseline | ActivityNet Captions · ActivityNet Captions | R@5 IoU=0.5: 74.87 -> 74.13 | No clear effect |
| Decouple query words using PoS into relation and modified features (ablation) | DeNet language encoding · Charades-STA | R@1 IoU=0.5: 57.47 -> 59.70 | Helped |
| Encode the modified feature, rather than the relation feature, as a Gaussian distribution (ablation) | DeNet language encoding · Charades-STA | R@(5,5) IoU=0.5: 54.32 -> 56.30 | Helped |
| Encode both relation and modified features as Gaussian distributions (ablation) | DeNet language encoding · Charades-STA | R@(5,5) IoU=0.5: 56.30 -> 46.20 | Hurt |
| Add the center-width auxiliary prediction head to boundary prediction (ablation) | DeNet temporal regression · Charades-STA | R@1 IoU=0.5: 57.88 -> 59.70 | Helped |
| Add the start-end boundary head to center-width prediction (ablation) | DeNet temporal regression · Charades-STA | R@1 IoU=0.5: 57.85 -> 59.70 | Helped |
| Use two independent single-output and multi-output branches (ablation) | DeNet temporal regression · Charades-STA | R@(5,5) IoU=0.5: 55.38 -> 56.30 | Helped |
| Use min-loss for the multi-output branch instead of single-output loss (ablation) | DeNet temporal regression · Charades-STA | R@(5,5) IoU=0.5: 42.18 -> 56.30 | Helped |
Closest papers by meaning
- Local-Global Video-Text Interactions for Temporal Grounding (CVPR 2020 · similarity 0.89)
- Universal Video Temporal Grounding with Generative Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.87)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.87)
- Comprehensive Visual Grounding for Video Description (AAAI 2024 · similarity 0.86)
- Multi-Pair Temporal Sentence Grounding via Multi-Thread Knowledge Transfer Network (AAAI 2025 · similarity 0.86)
- Bias-Conflict Sample Synthesis and Adversarial Removal Debias Strategy for Temporal Sentence Grounding in Video (AAAI 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.