Research map / Architectures and efficiency / Attention and transformers
Handwritten Mathematical Expression Recognition via Attention Aggregation Based Bi-directional Mutual Learning
Xiaohang Bian, Bo Qin, Xiaozhe Xin, Jianwu Li, Xuefeng Su, Yanfeng Wang
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 attention aggregation (AAM) to the DWAP baseline (ablation) | DWAP · CROHME 2014 | ExpRate: 50.60 to 52.64 | Improved |
| add bi-directional mutual learning (BML) to the DWAP baseline (ablation) | DWAP · CROHME 2014 | ExpRate: 50.60% to 55.23% | Improved |
| add both AAM and BML (ablation) | DWAP · CROHME 2014 | ExpRate: 50.60% to 56.85% | Improved |
| add AAM and BML to the L2R model and evaluate prefix and suffix recognition (ablation) | Uni-L2R · CROHME 2014, Prefix-2 | accuracy: 85.29% to 88.73% | Improved |
| add AAM and BML to the L2R model and evaluate suffix recognition (ablation) | Uni-L2R · CROHME 2014, Suffix-2 | accuracy: 80.02% to 84.37% | Improved |
| use bi-directional mutual learning rather than uni-directional mutual learning with AAM (ablation) | AUM · CROHME 2014 | ExpRate: 56.24 to 56.85 | Improved |
| apply AAM and BML to the LSTM decoder (ablation) | DWAP-LSTM · CROHME 2014 | ExpRate: 49.64% to 55.13% | Improved |
| apply AAM and BML to BTTR (ablation) | BTTR · CROHME 2014 | ExpRate: 48.13% to 51.47% | Improved |
| add AAM and BML to the L2R branch and assess long-expression recognition (ablation) | DWAP · CROHME 2014, sequence lengths [41,∞] | accuracy | Improved |
Closest papers by meaning
- Syntax-Aware Network for Handwritten Mathematical Expression Recognition (CVPR 2022 · similarity 0.89)
- SSAN: A Symbol Spatial-Aware Network for Handwritten Mathematical Expression Recognition (AAAI 2025 · similarity 0.88)
- Uni-MuMER: Unified Multi-Task Fine-Tuning of Vision-Language Model for Handwritten Mathematical Expression Recognition (NeurIPS 2025 · similarity 0.88)
- TAMER: Tree-Aware Transformer for Handwritten Mathematical Expression Recognition (AAAI 2025 · similarity 0.86)
- Generating Handwritten Mathematical Expressions From Symbol Graphs: An End-to-End Pipeline (CVPR 2024 · similarity 0.84)
- MetaWriter: Personalized Handwritten Text Recognition Using Meta-Learned Prompt Tuning (CVPR 2025 · similarity 0.81)
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.
