Research map / Video / Video-language understanding
Uncertainty-aware sign language video retrieval with probability distribution modeling
Xuan Wu, Hongxiang Li, yuanjiang luo, Xuxin Cheng, Xianwei Zhuang, Meng Cao, Keren Fu
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 optimal transport only to baseline (ablation) | CiCo baseline · How2Sign (T2V) | R@1: 56.4 -> 57.8 | Improved |
| add distribution modeling and random sampling without OT (ablation) | CiCo baseline · How2Sign (T2V) | R@1: 56.4 -> 56.8 | Improved |
| combine DM, RS and OT (ablation) | UPRet · How2Sign (T2V) | R@1: 58.4 -> 59.1 | Improved |
| model both text and video as distributions rather than only one (ablation) | UPRet · How2Sign (T2V) | R@1: 57.9 -> 59.1 (text only); 58.3 -> 59.1 (video only) | Improved |
| use MLP alone as distribution head (ablation) | UPRet · How2Sign (T2V) | R@1: 56.4 -> 55.2 | Hurt |
| combine MLP and attention in distribution head (ablation) | UPRet · How2Sign (T2V) | R@1: 57.9 -> 59.1 | Improved |
| increase Monte Carlo samples from 2 to 4, 6 or 8 (ablation) | UPRet · How2Sign | retrieval performance | Hurt |
| use a single Monte Carlo sample instead of 2 (ablation) | UPRet · How2Sign | retrieval performance | Hurt |
| add DM and OT modules to the baseline | CiCo baseline · How2Sign (T2V) | training iteration time: 0.63 -> 0.70 | Hurt |
| UPRet vs CiCo* on CSL-Daily mean rank | UPRet · CSL-Daily (T2V) | MnR: 5.8 -> 6.7 | Hurt |
Closest papers by meaning
- YouTube-SL-25: A Large-Scale, Open-Domain Multilingual Sign Language Parallel Corpus (ICLR 2025 · similarity 0.86)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.86)
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.85)
- Conditional Variational Autoencoder for Sign Language Translation with Cross-Modal Alignment (AAAI 2024 · similarity 0.85)
- Advanced Sign Language Video Generation with Compressed and Quantized Multi-Condition Tokenization (NeurIPS 2025 · similarity 0.85)
- KD-MSLRT: Lightweight Sign Language Recognition Model Based on Mediapipe and 3D to 1D Knowledge Distillation (AAAI 2025 · 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.
