Procedural Text Mining with Large Language Models

TL;DR

This paper explores using GPT-4 and in-context learning techniques to extract procedural information from unstructured PDF texts, demonstrating promising results and addressing data scarcity challenges in NLP procedure extraction.

Contribution

It introduces a novel application of large language models with in-context learning for procedural text mining from PDFs, highlighting customizations that improve extraction without extensive training data.

Findings

In-context learning enhances procedure extraction accuracy.

Ontology-based definitions improve model understanding.

Limited samples suffice for effective zero-shot learning.

Abstract

Recent advancements in the field of Natural Language Processing, particularly the development of large-scale language models that are pretrained on vast amounts of knowledge, are creating novel opportunities within the realm of Knowledge Engineering. In this paper, we investigate the usage of large language models (LLMs) in both zero-shot and in-context learning settings to tackle the problem of extracting procedures from unstructured PDF text in an incremental question-answering fashion. In particular, we leverage the current state-of-the-art GPT-4 (Generative Pre-trained Transformer 4) model, accompanied by two variations of in-context learning that involve an ontology with definitions of procedures and steps and a limited number of samples of few-shot learning. The findings highlight both the promise of this approach and the value of the in-context learning customisations. These modifications have the potential to significantly address the challenge of obtaining sufficient training data, a hurdle often encountered in deep learning-based Natural Language Processing techniques for procedure extraction.