Extracting information from scientific articles using natural language processing: towards the design of solid oxide fuel cells

• Grant number: 24/13738-2
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2024
• End date: August 31, 2025
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Amauri Jardim de Paula
• Grantee: André de Araújo Caetano
• Host Institution: Centro Nacional de Pesquisa em Energia e Materiais (CNPEM). Ministério da Ciência, Tecnologia e Inovação (Brasil). Campinas , SP, Brazil
• Associated research grant: 23/09820-2 - Materials by design: from quantum materials to energy applications, AP.TEM

Abstract

The 21st century has brought complex challenges to the field of materials science, but with notable prospects. We are living in a period of transformation driven by a growing global demand for sustainable technologies and innovative solutions. The evolution of new materials not only propels progress but also results in innovation across various sectors, including electronics, aerospace, construction, and healthcare. Historically, the search for innovative materials and their practical applications has long depended on laboratory methods following Thomas Edison's trial-and-error approach, involving extensive empirical experimental research and substantial resource utilization. However, the landscape of materials research has been significantly redefined by the profound impact of computational tools and, most notably, by the integration of machine learning (ML) models. This fusion of ML with materials science presents a remarkable opportunity to streamline the exploration, advancement, and fine-tuning of materials with an unparalleled degree of precision and efficiency. However, well-curated and accurate databases with experimental data to serve as input for training ML models are lacking. In this context, natural language processing (NLP) tools have emerged as the most efficient solution for extracting relevant information from the scientific literature and structuring it into databases. In the area of fuel cells, there are tens of thousands of scientific articles with information that, in theory, could train statistical models (ML) that would enable the establishment of a design platform for these materials. This platform could accelerate, for example, the development of efficient and durable catalysts. Thus, this project aims to explore NLP tools to process, analyze, and extract information from scientific articles in the field of materials for fuel cells. By bridging the gap between the vast corpus of scientific literature and practical knowledge, we aim to accelerate the discovery and advancement of materials for fuel cells, thereby contributing to the clean energy revolution.