Label Semantics for Few Shot Named Entity Recognition

TL;DR

This paper introduces a neural approach for few-shot named entity recognition that utilizes label semantics via dual BERT encoders, significantly improving performance in low-resource scenarios.

Contribution

It proposes a novel architecture that encodes label semantics to enhance few-shot NER, achieving state-of-the-art results and robustness in low-resource settings.

Findings

Improved few-shot NER performance across multiple benchmarks.

Effective use of label semantics boosts low-resource NER accuracy.

Achieves on-par results with standard benchmarks.

Abstract

We study the problem of few shot learning for named entity recognition. Specifically, we leverage the semantic information in the names of the labels as a way of giving the model additional signal and enriched priors. We propose a neural architecture that consists of two BERT encoders, one to encode the document and its tokens and another one to encode each of the labels in natural language format. Our model learns to match the representations of named entities computed by the first encoder with label representations computed by the second encoder. The label semantics signal is shown to support improved state-of-the-art results in multiple few shot NER benchmarks and on-par performance in standard benchmarks. Our model is especially effective in low resource settings.