Research map / Large language models / Language, safety and interpretability
Natural Language-Assisted Sign Language Recognition
Read the paperCheck 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 |
|---|---|---|---|
| Use NLA-SLR instead of SignBERT | MSASL1000 · MSASL1000 test | per-instance Top-1: 71.24 -> 72.56 | Helped |
| Use NLA-SLR instead of SAM-SLR-v2 (5-crop) | WLASL2000 · WLASL2000 test | per-instance Top-1: 59.39 -> 61.05 | Helped |
| Use NLA-SLR instead of BEST | NMFs-CSL · NMFs-CSL test | Top-1: 79.2 -> 83.4 | Helped |
| Add language-aware label smoothing (ablation) | VKNet on WLASL2000 · WLASL2000 test | per-instance Top-1: 57.19 -> 58.41 | Helped |
| Add sign mixup (intra- and inter-modality) (ablation) | VKNet on WLASL2000 · WLASL2000 test | per-instance Top-1: 57.19 -> 60.32 | Helped |
| Use both 32- and 64-frame video-keypoint inputs (ablation) | VKNet on WLASL2000 · WLASL2000 test | per-instance Top-1: 56.95 -> 57.19 | Helped |
| Add intra-modality mixup (ablation) | VKNet + Lang-LS on WLASL2000 · WLASL2000 test | per-instance Top-1: 58.41 -> 59.56 | Helped |
| Add inter-modality mixup (ablation) | VKNet + Lang-LS on WLASL2000 · WLASL2000 test | per-instance Top-1: 58.41 -> 59.66 | Helped |
| Combine intra- and inter-modality mixup (ablation) | VKNet + Lang-LS on WLASL2000 · WLASL2000 test | per-instance Top-1: 59.66 -> 61.05 | Helped |
| Integrate the auxiliary classifier FC2 into FC1 (ablation) | Inter-modality mixup on WLASL2000 · WLASL2000 test | per-instance Top-1: 59.87 -> 60.84 | Helped |
| Decay the inter-modality mixup loss weight (ablation) | Inter-modality mixup on WLASL2000 · WLASL2000 test | per-instance Top-1: 60.84 -> 61.05 | Helped |
| Use semantic rather than uniform label-smoothing weights (ablation) | Label smoothing on WLASL2000, ε=0.2 · WLASL2000 test | per-instance Top-1: 60.11 -> 61.05 | Helped |
Closest papers by meaning
- Improving Gloss-free Sign Language Translation by Reducing Representation Density (NeurIPS 2024 · similarity 0.89)
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation (NeurIPS 2025 · similarity 0.89)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.89)
- Conditional Variational Autoencoder for Sign Language Translation with Cross-Modal Alignment (AAAI 2024 · similarity 0.89)
- VSNet: Focusing on the Linguistic Characteristics of Sign Language (CVPR 2025 · similarity 0.88)
- Sign2GPT: Leveraging Large Language Models for Gloss-Free Sign Language Translation (ICLR 2024 · similarity 0.88)
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.