Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Simple Yet Effective Subsequence-Enhanced Approach for Cross-Domain NER
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 subsequence features with BMRU and fuse them with global BERT features using ACU | Cross-domain NER · POLITICS, SCIENCE, MUSIC, LITERATURE, AI | F1 | Helped |
| Continue pre-training on target-domain text using DAPT | Ours · POLITICS | F1: 68.65 -> 71.17 | Helped |
| Continue pre-training on target-domain text using DAPT | Ours · SCIENCE | F1: 72.42 -> 79.28 | Helped |
| Continue pre-training on target-domain text using DAPT | Ours · MUSIC | F1: 67.05 -> 69.22 | Helped |
| Continue pre-training on target-domain text using DAPT | Ours · LITERATURE | F1: 60.89 -> 63.79 | Helped |
| Use BMRU rather than GRU to extract subsequence features (ablation) | Cross-domain NER · CoNLL03 | F1: 69.60 -> 71.31 | Helped |
| Use BMRU rather than LSTM to extract subsequence features (ablation) | Cross-domain NER · MUSIC | F1: 66.20 -> 67.05 | Helped |
| Add memory states to subsequence feature extraction (ablation) | BERT-based cross-domain NER · CrossNER datasets | F1 | Helped |
| Use bidirectional rather than single-direction subsequence features (ablation) | BERT-based cross-domain NER · CrossNER datasets | F1 | Mixed |
| Use ACU rather than addition to combine local and global features (ablation) | BMRU + BERT · CoNLL03 | F1: 70.25 -> 71.31 | Helped |
| Use ACU rather than concatenation to combine local and global features (ablation) | BMRU + BERT · SCIENCE | F1: 71.96 -> 72.42 | Helped |
| Set subsequence radius k to 3 (subsequence length 7) (ablation) | Ours · MUSIC | F1 | Helped |
Closest papers by meaning
- Three Heads Are Better than One: Improving Cross-Domain NER with Progressive Decomposed Network (AAAI 2024 · similarity 0.92)
- MCL-NER: Cross-Lingual Named Entity Recognition via Multi-View Contrastive Learning (AAAI 2024 · similarity 0.89)
- UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition (ICLR 2024 · similarity 0.88)
- ConsistNER: Towards Instructive NER Demonstrations for LLMs with the Consistency of Ontology and Context (AAAI 2024 · similarity 0.88)
- PaDeLLM-NER: Parallel Decoding in Large Language Models for Named Entity Recognition (NeurIPS 2024 · similarity 0.87)
- Small Language Model Makes an Effective Long Text Extractor (AAAI 2025 · similarity 0.87)
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.