Research map / Video / Video-language understanding
Interventional Video Grounding With Dual Contrastive Learning
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 |
|---|---|---|---|
| Add IVG causal intervention using role, action, and object event proxies for confounders (ablation) | IVG-DCL · Charades-STA | mIoU: 45.34 -> 48.02 | Helped |
| Add query-video contrastive learning (QV-CL) (ablation) | IVG-DCL · Charades-STA | mIoU: 47.15 -> 48.02 | Helped |
| Add video-video contrastive learning (VV-CL) (ablation) | IVG-DCL · Charades-STA | mIoU: 47.24 -> 48.02 | Helped |
| Add dual contrastive learning (QV-CL and VV-CL) (ablation) | IVG-DCL · Charades-STA | mIoU: 46.94 -> 48.02 | Helped |
| Add both IVG and DCL rather than removing both components (ablation) | IVG-DCL · Charades-STA | mIoU: 42.11 -> 48.02 | Helped |
| Use alpha=1, beta=1 for the contrastive loss weights instead of alpha=0.1, beta=0.01 (ablation) | IVG-DCL · Charades-STA | mIoU: 48.02 -> 43.65 | Hurt |
| Use alpha=0.5, beta=0.1 instead of alpha=0.1, beta=0.01 (ablation) | IVG-DCL · Charades-STA | mIoU: 48.02 -> 46.09 | Hurt |
| Use alpha=0.1, beta=0.5 instead of alpha=0.1, beta=0.01 (ablation) | IVG-DCL · Charades-STA | mIoU: 48.02 -> 44.08 | Hurt |
| Use alpha=1.5, beta=1 instead of alpha=0.1, beta=0.01 (ablation) | IVG-DCL · Charades-STA | mIoU: 48.02 -> 44.89 | Hurt |
| Replace VSLNet with IVG-DCL | video grounding · Charades-STA | mIoU: 45.15 -> 48.02 | Helped |
| Replace VSLNet with IVG-DCL | video grounding · TACoS | mIoU: 24.11 -> 28.26 | Helped |
| Replace VSLNet with IVG-DCL | video grounding · ActivityNet Caption | mIoU: 43.19 -> 44.21 | Helped |
Closest papers by meaning
- Cross-modal Causal Relation Alignment for Video Question Grounding (CVPR 2025 · similarity 0.88)
- Multi-Scale Contrastive Learning for Video Temporal Grounding (AAAI 2025 · similarity 0.88)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.87)
- Time-R1: Post-Training Large Vision Language Model for Temporal Video Grounding (NeurIPS 2025 · similarity 0.87)
- Uncertainty-quantified Rollout Policy Adaptation for Unlabelled Cross-domain Video Temporal Grounding (NeurIPS 2025 · similarity 0.87)
- Learning Multi-Scale Video-Text Correspondence for Weakly Supervised Temporal Article Gronding (AAAI 2024 · similarity 0.87)
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.