Research map / Label-efficient and robust learning / Recognition and long-tail
Glyce: Glyph-vectors for Chinese Character Representations
Yuxian Meng, Wei Wu, Fei Wang, Xiaoya Li, Ping Nie, Fan Yin, Muyu Li, Qinghong Han, Xiaofei Sun, Jiwei Li
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 Glyce glyph embeddings | BERT · MSRA NER | F1: +0.74 | Improved |
| add Glyce glyph embeddings | BERT · Weibo NER | F1: +0.27 | Improved |
| add Glyce glyph embeddings | BERT · UD1 POS | F1: +1.35 | Improved |
| add Glyce glyph embeddings | Lattice-LSTM · MSRA NER | F1: +0.71 | Improved |
| add Glyce glyph embeddings | BERT · LCQMC | accuracy: 87.5 -> 88.7 | Improved |
| add Glyce embeddings | Biaffine parser · CTB5 dependency parsing | UAS: +0.9 | Improved |
| staged BERT fine-tune then glyph then joint training instead of direct joint (ablation) | Glyce+BERT · LCQMC | F1: 87.8 -> 88.8 | Improved |
| transformer output layer instead of BiLSTM (ablation) | Glyce+BERT · LCQMC | F1: 87.9 -> 88.8 | Improved |
| add auxiliary image-classification objective (ablation) | Glyce+BERT · LCQMC | F1: 88.4 -> 88.8 | Improved |
| Tianzige-CNN instead of vanilla CNN (ablation) | Glyce+BERT · LCQMC | F1: 87.4 -> 88.8 | Improved |
Closest papers by meaning
- GlyphMastero: A Glyph Encoder for High-Fidelity Scene Text Editing (CVPR 2025 · similarity 0.80)
- GlyphSR: A Simple Glyph-Aware Framework for Scene Text Image Super-Resolution (AAAI 2025 · similarity 0.78)
- DeepCalliFont: Few-Shot Chinese Calligraphy Font Synthesis by Integrating Dual-Modality Generative Models (AAAI 2024 · similarity 0.77)
- VecGlypher: Unified Vector Glyph Generation with Language Models (CVPR 2026 · similarity 0.77)
- GlyphDraw2: Automatic Generation of Complex Glyph Posters with Diffusion Models and Large Language Models (AAAI 2025 · similarity 0.76)
- GlyphAnchor: Enhancing Visual Text Rendering via Position-Anchored Glyph Priors (NeurIPS 2026 · similarity 0.75)
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.
