Research map / Probabilistic and scientific modeling / Molecules and proteins
Nested Named Entity Recognition with Partially-Observed TreeCRFs
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 |
|---|---|---|---|
| Use more latent labels (ablation) | PO-TreeCRF · ACE2004, ACE2005, and GENIA | precision and recall: ACE2004 (1 to 6): P 86.7 to 84.9, R 86.5 to 87.4; ACE2005 (1 to 6): P 84.5 to 82.6, R 86.4 to 87.8 | Mixed |
| Increase latent-state count beyond the number of observed states (ablation) | PO-TreeCRF · ACE2004, ACE2005, and GENIA | F1: ACE2004: 86.6 (1) to 86.1 (6); ACE2005: 85.4 (1) to 85.1 (6); GENIA: 78.2 (1) to 77.6 (4) | Hurt |
| Use 2 latent labels instead of 1 (ablation) | PO-TreeCRF · GENIA | F1: 78.2 to 78.2 | No clear effect |
| Use 3 latent labels instead of 1 (ablation) | PO-TreeCRF · ACE2004 | F1: 86.6 to 86.6 | No clear effect |
| Set latent-label count to 2 × the observed label size (ablation) | PO-TreeCRF · ACE2004, ACE2005, and GENIA | F1: ACE2004: 86.6 to 85.8; ACE2005: 85.4 to 84.7; GENIA: 78.2 to 76.9 | Hurt |
| Use Masked Inside for batchified training | PO-TreeCRF partial marginalization · ACE2004, one training epoch | GPU training time: 8m10s -> 3m20s | Helped |
| Infer trees using the NER objective rather than a syntax-tree objective | PO-TreeCRF · Training set, syntactical constituency-tree alignment | UAS: 10.83 vs 16.95; 5.61 vs 16.95; 0.15 vs 16.95 | Helped |
Closest papers by meaning
- Span Graph Transformer for Document-Level Named Entity Recognition (AAAI 2024 · similarity 0.87)
- Small Language Model Makes an Effective Long Text Extractor (AAAI 2025 · similarity 0.86)
- UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition (ICLR 2024 · similarity 0.86)
- PaDeLLM-NER: Parallel Decoding in Large Language Models for Named Entity Recognition (NeurIPS 2024 · similarity 0.85)
- MCL-NER: Cross-Lingual Named Entity Recognition via Multi-View Contrastive Learning (AAAI 2024 · similarity 0.84)
- ConsistNER: Towards Instructive NER Demonstrations for LLMs with the Consistency of Ontology and Context (AAAI 2024 · similarity 0.84)
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.