Evaluating Large Language Models for Structured Science Summarization in the Open Research Knowledge Graph

TL;DR

This study assesses the capability of large language models to automatically generate structured research summaries for the Open Research Knowledge Graph, comparing their performance to manual curation across multiple evaluation methods.

Contribution

It provides a comprehensive analysis of LLMs' effectiveness in producing structured scientific properties, highlighting their potential and need for further fine-tuning.

Findings

LLMs show potential in recommending structured science properties.

Performance varies across different evaluation metrics.

Further fine-tuning improves LLM alignment with scientific curation.

Abstract

Structured science summaries or research contributions using properties or dimensions beyond traditional keywords enhances science findability. Current methods, such as those used by the Open Research Knowledge Graph (ORKG), involve manually curating properties to describe research papers' contributions in a structured manner, but this is labor-intensive and inconsistent between the domain expert human curators. We propose using Large Language Models (LLMs) to automatically suggest these properties. However, it's essential to assess the readiness of LLMs like GPT-3.5, Llama 2, and Mistral for this task before application. Our study performs a comprehensive comparative analysis between ORKG's manually curated properties and those generated by the aforementioned state-of-the-art LLMs. We evaluate LLM performance through four unique perspectives: semantic alignment and deviation with ORKG properties, fine-grained properties mapping accuracy, SciNCL embeddings-based cosine similarity, and expert surveys comparing manual annotations with LLM outputs. These evaluations occur within a multidisciplinary science setting. Overall, LLMs show potential as recommendation systems for structuring science, but further finetuning is recommended to improve their alignment with scientific tasks and mimicry of human expertise.