Research map / Video / Video-language understanding
Rethinking Video-Language Model from the Language Input Perspective
Xiang Fang, Wanlong Fang, Changshuo Wang, Xiaoye Qu, Daizong Liu
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 |
|---|---|---|---|
| augment query texts (no plug-in) | X-Pool · MSR-VTT text-to-video | R@1: 46.9 -> 40.1 | Hurt |
| add the plug-and-play framework | X-Pool with augmented text · MSR-VTT text-to-video | R@1: 40.1 -> 45.9 | Improved |
| add the plug-and-play framework | T-MASS with augmented text · MSR-VTT text-to-video | R@1: 42.1 -> 51.4 | Improved |
| add the plug-and-play framework | T-MASS without augmentation · MSR-VTT text-to-video | R@1: 50.2 -> 52.3 | Improved |
| add the plug-and-play framework | BLIP-2 with augmented text · STAR Interaction | accuracy: 60.9 -> 66.2 | Improved |
| add the plug-and-play framework | G2L with augmented text · ActivityNet Captions R@1 IoU=0.5 | R@1: 33.01 -> 46.86 | Improved |
| add multi-level text augmentation (ablation) | Ours (G2L) · ActivityNet Captions R@1 IoU=0.5 | R@1: 40.28 -> 46.86 | Improved |
| add attribute-based text reasoning (ablation) | Ours (G2L) · ActivityNet Captions R@1 IoU=0.5 | R@1: 42.03 -> 46.86 | Improved |
| add video-guided self-weighted cross-modal bridging (ablation) | Ours (G2L) · ActivityNet Captions R@1 IoU=0.5 | R@1: 43.86 -> 46.86 | Improved |
| use own augmentation instead of back-translation (ablation) | Ours (BLIP-2) · NExT-QA Temporal | accuracy: 62.8 -> 69.2 | Improved |
| use own augmentation instead of LLaMA-7B (ablation) | Ours (BLIP-2) · NExT-QA Temporal | accuracy: 66.8 -> 69.2 | Improved |
| add S1 significance weight (ablation) | Ours (BLIP-2) · STAR Interaction | accuracy: 64.8 -> 66.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- X-Pool — X-Pool: Cross-Modal Language-Video Attention for Text-Video Retrieval (CVPR 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Needle In A Video Haystack: A Scalable Synthetic Evaluator for Video MLLMs (ICLR 2025 · similarity 0.87)
- T2V-CompBench: A Comprehensive Benchmark for Compositional Text-to-video Generation (CVPR 2025 · similarity 0.87)
- Mimir: Improving Video Diffusion Models for Precise Text Understanding (CVPR 2025 · similarity 0.87)
- Vgent: Graph-based Retrieval-Reasoning-Augmented Generation For Long Video Understanding (NeurIPS 2025 · similarity 0.87)
- Leveraging Textual Memory and Key Frame Reasoning for Full Video Understanding Using Off-the-Shelf LLMs and VLMs (Student Abstract) (AAAI 2025 · similarity 0.86)
- Distilling Vision-Language Models on Millions of Videos (CVPR 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.
