Research map / Video / Video-language understanding
Curriculum Multi-Negative Augmentation for Debiased Video Grounding
Read the paperCheck 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 |
|---|---|---|---|
| Apply curriculum multi-negative augmentation | 2D-TAN · Charades-CD i.i.d. and o.o.d. | dR@1,IoU=0.7: 28.18 -> 34.47 (i.i.d.); 13.73 -> 21.38 (o.o.d.) | Helped |
| Apply curriculum multi-negative augmentation | 2D-TAN · ActivityNet-CD i.i.d. and o.o.d. | dR@1,IoU=0.5: 40.87 -> 41.67 (i.i.d.); 18.86 -> 20.78 (o.o.d.) | Helped |
| Add clip-level cross-video negative augmentation only (ablation) | 2D-TAN · Charades-CD i.i.d. | R@1,IoU=0.5: 49.33 -> 50.67 | Helped |
| Add video-level cross-video negative augmentation only (ablation) | 2D-TAN · Charades-CD i.i.d. | R@1,IoU=0.5: 49.33 -> 57.23 | Helped |
| Add self-shuffled negative augmentation with masks only (ablation) | 2D-TAN · Charades-CD i.i.d. | R@1,IoU=0.5: 49.33 -> 47.75 | Hurt |
| Add all three negative augmentations without curriculum (ablation) | 2D-TAN · Charades-CD i.i.d. | R@1,IoU=0.5: 49.33 -> 55.77 | Helped |
| Add the multi-stage curriculum to all three negative augmentations (ablation) | 2D-TAN with Multi-NA · Charades-CD i.i.d. | R@1,IoU=0.5: 55.77 -> 58.57 | Helped |
| Add the multi-stage curriculum to all three negative augmentations (ablation) | 2D-TAN with Multi-NA · Charades-CD o.o.d. | R@1,IoU=0.5: 43.39 -> 44.75 | Helped |
Closest papers by meaning
- SnAG: Scalable and Accurate Video Grounding (CVPR 2024 · similarity 0.87)
- Bias-Conflict Sample Synthesis and Adversarial Removal Debias Strategy for Temporal Sentence Grounding in Video (AAAI 2024 · similarity 0.87)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.86)
- Uncertainty-quantified Rollout Policy Adaptation for Unlabelled Cross-domain Video Temporal Grounding (NeurIPS 2025 · similarity 0.86)
- Siamese Learning with Joint Alignment and Regression for Weakly-Supervised Video Paragraph Grounding (CVPR 2024 · similarity 0.86)
- DeCafNet: Delegate and Conquer for Efficient Temporal Grounding in Long Videos (CVPR 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.