Research map / Large language models / Language, safety and interpretability
Deep Short Text Classification with Knowledge Powered Attention
Jindong Chen, Yizhou Hu, Jingping Liu, Yanghua Xiao, Haiyun Jiang
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 knowledge powered attention to the short-text classifier | KPCNN-style short text classification · Weibo, Topic, Product Review, News Title | accuracy: 0.4240 -> 0.4320; 0.8643 -> 0.8814; 0.7340 -> 0.7430; 0.7878 -> 0.8011 | Improved |
| combine C-ST and C-CS attention rather than use C-ST only (ablation) | STCKA · Weibo | accuracy: 0.4160 -> 0.4320 | Improved |
| combine C-ST and C-CS attention rather than use C-CS only (ablation) | STCKA · Topic | accuracy: 0.8600 -> 0.8786 | Improved |
| set γ to 0.25 instead of 0.50 (ablation) | STCKA · Weibo | accuracy: 0.4260 -> 0.4320 | Improved |
| set γ to 0.50 instead of 0.25 (ablation) | STCKA · Topic | accuracy: 0.8700 -> 0.8786 | Improved |
| fine-tune pretrained character and word embeddings rather than keep them static (ablation) | STCKA · Weibo, Topic, Product Review, News Title | accuracy: 0.4240 -> 0.4320; 0.8600 -> 0.8814; 0.7350 -> 0.7430; 0.7889 -> 0.8011 | Improved |
| use pretrained embeddings rather than random initialization (ablation) | STCKA · Weibo | accuracy: 0.3780 -> 0.4240 | Improved |
| use pretrained embeddings rather than random initialization (ablation) | STCKA · Topic | accuracy: 0.8414 -> 0.8600 | Improved |
| randomly initialize embeddings rather than keep pretrained embeddings static (ablation) | STCKA · News Title | accuracy: 0.7889 -> 0.7930 | Improved |
Closest papers by meaning
- Self-Attention Based Text Knowledge Mining for Text Detection (CVPR 2021 · similarity 0.86)
- Boosting Short Text Classification with Multi-Source Information Exploration and Dual-Level Contrastive Learning (AAAI 2025 · similarity 0.85)
- Weakly-Supervised Hierarchical Text Classification (AAAI 2019 · similarity 0.84)
- SALNet: Semi-supervised Few-Shot Text Classification with Attention-based Lexicon Construction (AAAI 2021 · similarity 0.84)
- Improved Graph Contrastive Learning for Short Text Classification (AAAI 2024 · similarity 0.83)
- Liberating Seen Classes: Boosting Few-Shot and Zero-Shot Text Classification via Anchor Generation and Classification Reframing (AAAI 2024 · similarity 0.83)
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.
