Research map / Large language models / Language, safety and interpretability
Q-BridgeNet: A Quantization Network for Cross-Lingual Sign Language Translation
Liqian Feng, Lintao Wang, Xiaochen Liu, Anusha Withana, Ken-Tye Yong, Dehui Kong, Zhiyong Wang, Kun Hu
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 language-specific codebooks with a shared base instead of one global codebook (ablation) | Q-BridgeNet · PHOENIX14T | BLEU-4: 21.38 -> 33.40 | Improved |
| add shared base codebook (ablation) | Q-BridgeNet · PHOENIX14T | BLEU-4: 22.62 -> 33.40 | Improved |
| use residual stacking instead of flat parallel codebooks (ablation) | Q-BridgeNet · PHOENIX14T | BLEU-4: 27.32 -> 33.40 | Improved |
| add alternating segmentation and quantization refinement (ablation) | Q-BridgeNet · PHOENIX14T | BLEU-4: 28.61 -> 33.40 | Improved |
| add alternating refinement (ablation) | Q-BridgeNet · How2Sign | BLEU-4: 13.39 -> 16.44 | Improved |
| add shared base codebook (ablation) | Q-BridgeNet · How2Sign | BLEU-4: 9.29 -> 16.44 | Improved |
| add residual stack (ablation) | Q-BridgeNet · CSL-Daily | BLEU-4: 21.49 -> 26.33 | Improved |
| train multilingual instead of monolingual on native pair | Q-BridgeNet DGS-DE · PHOENIX14T | BLEU-4: 30.16 -> 33.62 | Improved |
| train multilingual instead of monolingual on native pair | Q-BridgeNet ASL-EN · How2Sign | BLEU-4: 14.35 -> 16.55 | Improved |
| train multilingual instead of monolingual on native pair | Q-BridgeNet CSL-ZH · CSL-Daily | BLEU-4: 24.08 -> 26.91 | Improved |
| use cross-lingual sign modeling instead of monolingual SLT then GPT translation (ablation) | Q-BridgeNet · CSL-Daily | BLEU-4: 23.18 -> 26.33 | Improved |
| use GPT-5 instead of GPT-4.1 for caption translation (ablation) | Q-BridgeNet · PHOENIX14T | BLEU-4: 33.07 -> 33.40 | Improved |
Closest papers by meaning
- Scaling Sign Language Translation (NeurIPS 2024 · similarity 0.91)
- LLMs are Good Sign Language Translators (CVPR 2024 · similarity 0.91)
- Lost in Translation, Found in Embeddings: Sign Language Translation and Alignment (NeurIPS 2026 · similarity 0.90)
- KD-MSLRT: Lightweight Sign Language Recognition Model Based on Mediapipe and 3D to 1D Knowledge Distillation (AAAI 2025 · similarity 0.90)
- Lost in Translation, Found in Context: Sign Language Translation with Contextual Cues (CVPR 2025 · similarity 0.90)
- Conditional Variational Autoencoder for Sign Language Translation with Cross-Modal Alignment (AAAI 2024 · similarity 0.89)
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.
