Research map / Video / Video-language understanding
Weakly Supervised Video Moment Localization with Contrastive Negative Sample Mining
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 |
|---|---|---|---|
| use CNM instead of CRM | weakly supervised video moment localization · ActivityNet Captions val 2 | Recall@IoU=0.1: 81.61 -> 78.13 | Hurt |
| use CNM instead of WSTAN | weakly supervised video moment localization · ActivityNet Captions val 2 | Recall@IoU=0.3: 52.45 -> 55.68 | Helped |
| use CNM instead of CRM | weakly supervised video moment localization · ActivityNet Captions val 2 | Recall@IoU=0.5: 32.19 -> 33.33 | Helped |
| use CNM instead of RTBPN | weakly supervised video moment localization · Charades-STA test | Recall@IoU=0.3: 60.04 -> 60.04 | No clear effect |
| use CNM instead of LCNet | weakly supervised video moment localization · Charades-STA test | Recall@IoU=0.5: 39.19 -> 35.15 | Hurt |
| use CNM instead of LCNet | weakly supervised video moment localization · Charades-STA test | Recall@IoU=0.7: 18.87 -> 14.95 | Hurt |
| add a learnable Gaussian mask generator instead of sliding windows and policy gradient (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.3: 47.71 -> 55.68 | Helped |
| add a learnable Gaussian mask generator instead of sliding windows and policy gradient (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.5: 26.98 -> 33.33 | Helped |
| add hard and easy intra-video negatives (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.3: 40.26 -> 55.68 | Helped |
| add hard and easy intra-video negatives (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.5: 24.93 -> 33.33 | Helped |
| add hard negative samples (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.5: 31.40 -> 33.33 | Helped |
| add easy negative samples (ablation) | CNM · ActivityNet Captions val 2 | Recall@IoU=0.5: 30.94 -> 33.33 | Helped |
Closest papers by meaning
- Weakly-Supervised Video Re-Localization with Multiscale Attention Model (AAAI 2020 · similarity 0.88)
- Weakly Supervised Temporal Sentence Grounding With Gaussian-Based Contrastive Proposal Learning (CVPR 2022 · similarity 0.88)
- Negative Sample Matters: A Renaissance of Metric Learning for Temporal Grounding (AAAI 2022 · similarity 0.88)
- Omnipotent Distillation with LLMs for Weakly-Supervised Natural Language Video Localization: When Divergence Meets Consistency (AAAI 2024 · similarity 0.87)
- Multi-Scale Contrastive Learning for Video Temporal Grounding (AAAI 2025 · similarity 0.87)
- Transferable Video Moment Localization by Moment-Guided Query Prompting (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.