SEDA: Enhancing Discontinuous NER with Self-Adapted Data Augmentation
Analysis
The paper introduces SEDA, a novel data augmentation technique specifically designed to improve grid-based discontinuous Named Entity Recognition (NER) models. This targeted approach suggests a potential for significant performance gains in complex NER tasks.
Key Takeaways
- •SEDA is a novel data augmentation method for discontinuous NER.
- •It's designed to boost grid-based NER models.
- •The approach is entity-centric, implying a focus on improving entity recognition accuracy.
Reference / Citation
View Original"SEDA is a self-adapted entity-centric data augmentation technique."