Few-NERD: A Few-Shot Named Entity Recognition Dataset

TL;DR

Few-NERD introduces the first large-scale, hierarchical, human-annotated few-shot NER dataset, highlighting the challenges of recognizing fine-grained entity types in a few-shot setting and providing a comprehensive benchmark for future research.

Contribution

This paper presents Few-NERD, the largest human-crafted few-shot NER dataset with hierarchical entity types, enabling more realistic and challenging evaluation of NER models.

Findings

Few-NERD is more challenging than existing datasets.

Models struggle with fine-grained entity recognition in few-shot scenarios.

The dataset reveals the need for improved few-shot NER methods.

Abstract

Recently, considerable literature has grown up around the theme of few-shot named entity recognition (NER), but little published benchmark data specifically focused on the practical and challenging task. Current approaches collect existing supervised NER datasets and re-organize them to the few-shot setting for empirical study. These strategies conventionally aim to recognize coarse-grained entity types with few examples, while in practice, most unseen entity types are fine-grained. In this paper, we present Few-NERD, a large-scale human-annotated few-shot NER dataset with a hierarchy of 8 coarse-grained and 66 fine-grained entity types. Few-NERD consists of 188,238 sentences from Wikipedia, 4,601,160 words are included and each is annotated as context or a part of a two-level entity type. To the best of our knowledge, this is the first few-shot NER dataset and the largest human-crafted NER dataset. We construct benchmark tasks with different emphases to comprehensively assess the generalization capability of models. Extensive empirical results and analysis show that Few-NERD is challenging and the problem requires further research. We make Few-NERD public at https://ningding97.github.io/fewnerd/.