Research map / Video / Video-language understanding
Compositional Transformation Reasoning for Composed Video Retrieval
Sihong Huang, Jiaxin Wu, Dongmei Jiang, Yi Cai, Yaowei Wang, Xiaoyong Wei
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 |
|---|---|---|---|
| use pairwise relative comparison instead of absolute relevance scoring in MLLM (ablation) | MoRe · EgoCVR (global) | R@1: 7.4 -> 20.4 | Improved |
| replace original modification text with predicted fine-grained target semantics for retrieval (ablation) | MoRe · EgoCVR text-to-video | R@1: 0.9 -> 5.9 | Improved |
| replace original modification text with predicted fine-grained target semantics for retrieval (ablation) | MoRe · WebVid-CoVR text-to-video | R@1: 24.4 -> 47.6 | Improved |
| use predicted semantics in ROVCS with MLLM reasoning stage (ablation) | MoRe · EgoCVR (global) | R@1: 16.3 -> 20.4 | Improved |
| use predicted semantics in ROVCS with MLLM reasoning stage (ablation) | MoRe · WebVid-CoVR | R@1: 60.9 -> 63.0 | Improved |
| add MLLM pairwise reasoning replacing text-similarity stage 2 (with ROVCS) (ablation) | MoRe · EgoCVR (global) | R@1: 8.4 -> 20.4 | Improved |
| add ROVCS candidate selection over visual-similarity stage 1 (with MLLM reasoning) (ablation) | MoRe · EgoCVR (global) | R@1: 15.8 -> 20.4 | Improved |
| add ROVCS candidate selection over visual-similarity stage 1 (with MLLM reasoning) (ablation) | MoRe · WebVid-CoVR | R@1: 60.3 -> 63.0 | Improved |
| ROVCS alone (with text-similarity stage 2) vs TFR-CVR baseline (ablation) | MoRe · EgoCVR (global) | R@1: 14.1 -> 8.4 | Hurt |
| ROVCS candidate selection instead of visual similarity (ablation) | MoRe · EgoCVR target-in-candidate ratio | target in candidate set (%): 64.5 -> 74.5 | Improved |
| reduce candidates K from 15 to 10 (ablation) | MoRe · EgoCVR (global) | R@1: 20.4 -> 17.2; runtime 23.9s -> 10.4s | Mixed |
| Swiss tournament instead of full pairwise comparison (ablation) | MoRe · EgoCVR (global) | R@1: 20.4 -> 17.7; runtime 23.9s -> 12.8s | Mixed |
Closest papers by meaning
- VELOCITI: Benchmarking Video-Language Compositional Reasoning with Strict Entailment (CVPR 2025 · similarity 0.88)
- Vgent: Graph-based Retrieval-Reasoning-Augmented Generation For Long Video Understanding (NeurIPS 2025 · similarity 0.88)
- From Play to Replay: Composed Video Retrieval for Temporally Fine-Grained Videos (NeurIPS 2025 · similarity 0.87)
- VideoComp: Advancing Fine-Grained Compositional and Temporal Alignment in Video-Text Models (CVPR 2025 · similarity 0.87)
- Video-RAG: Visually-aligned Retrieval-Augmented Long Video Comprehension (NeurIPS 2025 · similarity 0.87)
- Towards Balanced Alignment: Modal-Enhanced Semantic Modeling for Video Moment Retrieval (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.
