Research map / Video / Video-language understanding
Can Shuffling Video Benefit Temporal Bias Problem: A Novel Training Framework for Temporal Grounding
Jiachang Hao, Haifeng Sun, Pengfei Ren, Jingyu Wang, Qi Qi, Jianxin Liao
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 intra-video cross-modal matching loss (ablation) | Lg · Charades-CD test-ood | mIoU: 35.05 -> 42.66 | Improved |
| add intra-video cross-modal matching loss (ablation) | Lg · ActivityNet-CD test-ood | mIoU: 26.75 -> 29.77 | Improved |
| add inter-video consistency loss (ablation) | Lg · Charades-CD test-ood | mIoU: 35.05 -> 36.36 | Improved |
| add inter-video consistency loss (ablation) | Lg · ActivityNet-CD test-ood | mIoU: 26.75 -> 26.27 | Hurt |
| add temporal order discrimination loss alone (ablation) | Lg · Charades-CD test-ood | mIoU: 35.05 -> 30.47 | Hurt |
| add temporal order discrimination loss alone (ablation) | Lg · ActivityNet-CD test-ood | mIoU: 26.75 -> 26.43 | Hurt |
| add temporal order discrimination loss (ablation) | Lg+Lintra+Linter · Charades-CD test-ood | mIoU: 42.93 -> 44.30 | Improved |
| add temporal order discrimination loss (ablation) | Lg+Lintra+Linter · ActivityNet-CD test-ood | mIoU: 30.04 -> 30.45 | Improved |
| add inter-video loss to intra loss (ablation) | Lg+Lintra · Charades-CD test-ood | mIoU: 42.66 -> 42.93 | Improved |
| add shuffled videos as data augmentation (ablation) | Baseline · Charades-CD test-ood | mIoU: 39.52 -> 39.06 | Hurt |
| add shuffled videos as data augmentation (ablation) | Baseline · ActivityNet-CD test-ood | mIoU: 28.41 -> 29.59 | Improved |
| add CSMM to data augmentation (ablation) | Baseline+DataAug · ActivityNet-CD test-ood | mIoU: 29.59 -> 20.29 | Hurt |
Closest papers by meaning
- Training-free Video Temporal Grounding using Large-scale Pre-trained Models (ECCV 2024 · similarity 0.90)
- SnAG: Scalable and Accurate Video Grounding (CVPR 2024 · similarity 0.89)
- Universal Video Temporal Grounding with Generative Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.89)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.89)
- Multi-Pair Temporal Sentence Grounding via Multi-Thread Knowledge Transfer Network (AAAI 2025 · similarity 0.88)
- Bias-Conflict Sample Synthesis and Adversarial Removal Debias Strategy for Temporal Sentence Grounding in Video (AAAI 2024 · similarity 0.88)
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.
