Predicting New Research Directions in Materials Science using Large Language Models and Concept Graphs

TL;DR

This work investigates the use of large language models (LLMs) for the purpose of extracting the main concepts and semantic information from scientific abstracts in the domain of materials science to find links that were not noticed by humans and thus to suggest inspiring near/mid-term future research directions.

Abstract

Due to an exponential increase in published research articles, it is impossible for individual scientists to read all publications, even within their own research field. In this work, we investigate the use of large language models (LLMs) for the purpose of extracting the main concepts and semantic information from scientific abstracts in the domain of materials science to find links that were not noticed by humans and thus to suggest inspiring near/mid-term future research directions. We show that LLMs can extract concepts more efficiently than automated keyword extraction methods to build a concept graph as an abstraction of the scientific literature. A machine learning model is trained to predict emerging combinations of concepts, i.e. new research ideas, based on historical data. We demonstrate that integrating semantic concept information leads to an increased prediction performance. The applicability of our model is demonstrated in qualitative interviews with domain experts based on individualized model suggestions. We show that the model can inspire materials scientists in their creative thinking process by predicting innovative combinations of topics that have not yet been investigated.

Figures (8)

• Figure 1: Map of the materials science. UMAPmcinnes_umap_2018 2D projection of all extracted concepts with the highest-degree concepts in each square of length 2 highlighted and annotated ('Highest degree in tile'). Yellow and purple background colors respectively indicate high and low concept densities calculated using Kernel Density Estimation (KDE) is indicated by background color.
• Figure 2: Performance metrics (ROC and the respective AUC) for our link prediction models on the test set ($T_{\text{test}} = [2020, 2022]$). Markers highlight the performances at a threshold of $0.5$. (a) ROC curves on all data points with a zoomed-in view of the low false positive rate region in the inset. Panels (b) and (c) display the respective performance metrics for $d_{\text{prev}} = 2$ and $d_{\text{prev}} = 3$. Best result in bold.
• Figure 3: Overview of the report generation: (1) Selection of abstracts of recent publications, (2) Extraction of individual concepts using our fine-tuned LLM, and (3) Suggestion of combinations in a standardized report based on our best ML model for link prediction.
• Figure 4: Exfoliated graphene oxide (resulting in multilayer graphene) covered with 200nm thick iron oxide layers. a) shows the flake morphology of exfoliated layers, b) shows the iron oxide layer covering the whole surface.
• Figure 5: Overview of the link prediction workflow: (a) gathering of materials science abstracts with OpenAlex priem_openalex_2022, (b) extraction of concepts using LLMs, (c) creation of the semantics-aware-concept graph, and (d) prediction of new research directions.