Few-Shot Named Entity Recognition: An Empirical Baseline Study

Keyphrases

• Named Entity Recognition 100%
• Few-shot 100%
• Labeled Data 50%
• In-domain 50%
• Pre-trained Language Model 50%
• Learning Settings 25%
• Recognition Schemes 25%
• Recognition System 25%
• Linear Classifier 25%
• Domain Name 25%
• Model Generalization 25%
• Entity Typing 25%
• Self-training 25%
• Transformer-based 25%
• Recognition Dataset 25%
• Training-free 25%
• Few-shot Learning 25%
• Meta-learning 25%
• Noisy Web Data 25%
• Few-shot Setting 25%

Computer Science

• Named Entity Recognition 100%
• Pre-Trained Language Models 66%
• Linear Classifier 33%
• New-State 33%
• Data Domain 33%
• Generalization Ability 33%
• Named Entity Recognition System 33%
• Few-Shot Learning 33%