InstructionNER: A Multi-Task Instruction-Based Generative Framework for Few-shot NER

TL;DR

InstructionNER is a novel generative framework that reformulates few-shot NER as a generation task with instructions, auxiliary tasks, and achieves superior performance on multiple datasets.

Contribution

It introduces a multi-task instruction-based approach for low-resource NER, incorporating auxiliary tasks to improve boundary detection and type understanding.

Findings

Outperforms baselines on five datasets in few-shot scenarios.

Reformulates NER as a generation problem with instructions.

Uses auxiliary tasks to enhance entity boundary and type comprehension.

Abstract

Recently, prompt-based methods have achieved significant performance in few-shot learning scenarios by bridging the gap between language model pre-training and fine-tuning for downstream tasks. However, existing prompt templates are mostly designed for sentence-level tasks and are inappropriate for sequence labeling objectives. To address the above issue, we propose a multi-task instruction-based generative framework, named InstructionNER, for low-resource named entity recognition. Specifically, we reformulate the NER task as a generation problem, which enriches source sentences with task-specific instructions and answer options, then inferences the entities and types in natural language. We further propose two auxiliary tasks, including entity extraction and entity typing, which enable the model to capture more boundary information of entities and deepen the understanding of entity type semantics, respectively. Experimental results show that our method consistently outperforms other baselines on five datasets in few-shot settings.