Domain-Transferable Method for Named Entity Recognition Task

TL;DR

This paper introduces a domain-transferable NER method that learns domain-specific models without labeled data by leveraging unsupervised supervision and neural model interactions, reducing the need for costly annotations.

Contribution

It proposes a novel approach for domain-specific NER that requires no human-labeled data, enabling effective model training across diverse domains.

Findings

Method achieves competitive performance without labeled data

Neural models can learn from each other in unsupervised settings

Code, data, and models are publicly available

Abstract

Named Entity Recognition (NER) is a fundamental task in the fields of natural language processing and information extraction. NER has been widely used as a standalone tool or an essential component in a variety of applications such as question answering, dialogue assistants and knowledge graphs development. However, training reliable NER models requires a large amount of labelled data which is expensive to obtain, particularly in specialized domains. This paper describes a method to learn a domain-specific NER model for an arbitrary set of named entities when domain-specific supervision is not available. We assume that the supervision can be obtained with no human effort, and neural models can learn from each other. The code, data and models are publicly available.