Example-Based Named Entity Recognition
arXiv:2008.10570
Abstract
We present a novel approach to named entity recognition (NER) in the presence of scarce data that we call example-based NER. Our train-free few-shot learning approach takes inspiration from question-answering to identify entity spans in a new and unseen domain. In comparison with the current state-of-the-art, the proposed method performs significantly better, especially when using a low number of support examples.
15 pages, 6 figures, 5 tables with appendix
References in corpus (11)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Prototypical Networks for Few-shot Learning
- Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
- Language Models are Few-Shot Learners
- REALM: Retrieval-Augmented Language Model Pre-Training
- Few-shot classification in Named Entity Recognition Task
- A Unified MRC Framework for Named Entity Recognition
- Entity-Switched Datasets: An Approach to Auditing the In-Domain Robustness of Named Entity Recognition Models
- Coreference Resolution as Query-based Span Prediction
- A Rigorous Study on Named Entity Recognition: Can Fine-tuning Pretrained Model Lead to the Promised Land?
- Partially-Typed NER Datasets Integration: Connecting Practice to Theory
Cited by in corpus (6)
- Few-Shot Named Entity Recognition: A Comprehensive Study
- Neural Data Augmentation via Example Extrapolation
- Template-Based Named Entity Recognition Using BART
- ContrastNER: Contrastive-based Prompt Tuning for Few-shot NER
- Few-shot Intent Classification and Slot Filling with Retrieved Examples
- Learning from Language Description: Low-shot Named Entity Recognition via Decomposed Framework