UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition

TL;DR

UniversalNER demonstrates that targeted distillation from large language models can produce small, highly accurate open NER models that outperform larger instruction-tuned models across diverse domains without direct supervision.

Contribution

The paper introduces a targeted distillation approach for open NER, achieving state-of-the-art accuracy with small models by distilling from ChatGPT without supervision.

Findings

UniversalNER outperforms Alpaca and Vicuna by over 30 F1 points.

UniversalNER surpasses state-of-the-art multi-task systems like InstructUIE.

Achieves high accuracy across 43 datasets in diverse domains.

Abstract

Large language models (LLMs) have demonstrated remarkable generalizability, such as understanding arbitrary entities and relations. Instruction tuning has proven effective for distilling LLMs into more cost-efficient models such as Alpaca and Vicuna. Yet such student models still trail the original LLMs by large margins in downstream applications. In this paper, we explore targeted distillation with mission-focused instruction tuning to train student models that can excel in a broad application class such as open information extraction. Using named entity recognition (NER) for case study, we show how ChatGPT can be distilled into much smaller UniversalNER models for open NER. For evaluation, we assemble the largest NER benchmark to date, comprising 43 datasets across 9 diverse domains such as biomedicine, programming, social media, law, finance. Without using any direct supervision, UniversalNER attains remarkable NER accuracy across tens of thousands of entity types, outperforming general instruction-tuned models such as Alpaca and Vicuna by over 30 absolute F1 points in average. With a tiny fraction of parameters, UniversalNER not only acquires ChatGPT's capability in recognizing arbitrary entity types, but also outperforms its NER accuracy by 7-9 absolute F1 points in average. Remarkably, UniversalNER even outperforms by a large margin state-of-the-art multi-task instruction-tuned systems such as InstructUIE, which uses supervised NER examples. We also conduct thorough ablation studies to assess the impact of various components in our distillation approach. We release the distillation recipe, data, and UniversalNER models to facilitate future research on targeted distillation.