Research map / Video / Video-language understanding
Distribution-Alignment Bridge for Uncertainty-Aware Text-to-Video Retrieval
Kyeongmo Chae, Jihoon Lee, Sangtae Ahn
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 probabilistic embedding and distribution bridge to the X-Pool baseline (ablation) | DAB · MSR-VTT 5-epoch ablation | R@1: 46.9 -> 52.6 | Improved |
| add probabilistic embedding alone to baseline (ablation) | X-Pool · MSR-VTT 5-epoch ablation | MnR: 14.3 -> 7.4 | Improved |
| add probabilistic embedding alone to baseline (ablation) | X-Pool · MSR-VTT 5-epoch ablation | R@1: 46.9 -> 44.3 | Hurt |
| add the bridge alone to baseline (ablation) | X-Pool · MSR-VTT 5-epoch ablation | R@1: 46.9 -> 43.2 | Hurt |
| add the bridge on top of probabilistic embedding (ablation) | DAB · MSR-VTT 5-epoch ablation | R@1: 44.3 -> 52.6 | Improved |
| use directional KL cost instead of W2 (ablation) | DAB · MSR-VTT 5-epoch ablation | R@1: 46.8 -> 52.6 | Improved |
| use directional KL cost instead of W2 (ablation) | DAB · MSR-VTT 5-epoch ablation | MnR: 6.6 -> 5.3 | Improved |
| increase truncated steps T' from 8 to 32 (ablation) | DAB · MSR-VTT 5-epoch ablation | R@1: 49.0 -> 52.6 | Improved |
| increase truncated steps T' from 32 to 64 (ablation) | DAB · MSR-VTT 5-epoch ablation | R@1: 52.6 -> 51.9 | No clear effect |
| increase truncated steps T' from 32 to 64 (ablation) | DAB · MSR-VTT 5-epoch ablation | latency: 51.8 -> 68.9 | Hurt |
| replace DITS point-wise refinement with distribution alignment | DAB · MSR-VTT 1K-A (ViT-B/32) | R@1: 51.9 -> 56.2 | Improved |
| replace DITS with DAB | DAB · MSR-VTT 1K-A (ViT-B/32) | MnR: 11.6 -> 4.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- X-Pool — X-Pool: Cross-Modal Language-Video Attention for Text-Video Retrieval (CVPR 2022) · better on 4 of 4 shared results
- UATVR — UATVR: Uncertainty-Adaptive Text-Video Retrieval (ICCV 2023) · better on 2 of 2 shared results
- DiffusionRet — DiffusionRet: Generative Text-Video Retrieval with Diffusion Model (ICCV 2023) · better on 2 of 2 shared results
- Cap4Video — Cap4Video: What Can Auxiliary Captions Do for Text-Video Retrieval? (CVPR 2023) · better on 1 of 1 shared result
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
- Rebalancing Contrastive Alignment with Bottlenecked Semantic Increments in Text-Video Retrieval (NeurIPS 2025 · similarity 0.88)
- Rethinking Noisy Video-Text Retrieval via Relation-aware Alignment (CVPR 2025 · similarity 0.88)
- Mitigating Semantic Collapse in Partially Relevant Video Retrieval (NeurIPS 2025 · similarity 0.87)
- Diffusion-Inspired Truncated Sampler for Text-Video Retrieval (NeurIPS 2024 · similarity 0.87)
- Inference-Time Text-to-Video Alignment with Diffusion Latent Beam Search (NeurIPS 2025 · similarity 0.86)
- Text Is MASS: Modeling as Stochastic Embedding for Text-Video Retrieval (CVPR 2024 · similarity 0.85)
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.
