Research map / Architectures and efficiency / Attention and transformers
Mention and Entity Description Co-Attention for Entity Disambiguation
Feng Nie, Yunbo Cao, Jinpeng Wang, Chin-Yew Lin, Rong Pan
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 entity type to the co-attention affinity matrix (ablation) | CoAtt-Single · ACE | F1: 84.0 -> 84.9 | Improved |
| add entity type to the co-attention affinity matrix (ablation) | CoAtt-Single · KBP 2010 | F1: 82.4 -> 84.6 | Improved |
| add entity type to the co-attention affinity matrix (ablation) | CoAtt-Single · CoNLL | F1: 82.3 -> 83.1 | Improved |
| add co-attention on both context and description vs uniform attention (ablation) | LSTM-AVG · ACE, CoNLL, KBP 2010 | F1: 82.9 -> 84.0 (ACE), 81.6 -> 82.3 (CoNLL), 80.1 -> 82.4 (KBP) | Improved |
| add single-side attention (context only or description only) (ablation) | LSTM-AVG · ACE, CoNLL, KBP 2010 | F1: Context-ATT 83.5/80.8/80.5 vs LSTM-AVG 82.9/81.6/80.1 | No clear effect |
| treat entity type as an extra description sentence instead of type-aware affinity (ablation) | TypeCoAtt-Single · CoNLL | F1: 83.1 -> 81.9 | Hurt |
| use only entity type to represent entity (drop descriptions) (ablation) | TypeCoAtt-Single · ACE | F1: 84.9 -> 56.5 | Hurt |
| pre-train on 40,000 Wikipedia anchor queries | TypeCoAtt · KBP2010 | F1: 84.6 -> 86.4 | Improved |
| add sparse features to TypeCoAtt | TypeCoAtt (pre-train) · CoNLL | F1: 83.4 -> 89.8 | Improved |
Closest papers by meaning
- Entity Structure Within and Throughout: Modeling Mention Dependencies for Document-Level Relation Extraction (AAAI 2021 · similarity 0.87)
- A Supervised Multi-Head Self-Attention Network for Nested Named Entity Recognition (AAAI 2021 · similarity 0.87)
- Task-Specific Representation Learning for Web-Scale Entity Disambiguation (AAAI 2018 · similarity 0.87)
- GuideNER: Annotation Guidelines Are Better than Examples for In-Context Named Entity Recognition (AAAI 2025 · similarity 0.83)
- MCL-NER: Cross-Lingual Named Entity Recognition via Multi-View Contrastive Learning (AAAI 2024 · similarity 0.83)
- UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition (ICLR 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.
