Seminar: James Moraes de Almeida, Professor, Ilum Escola de Ciência

Materials informatics represents the powerful convergence of high-throughput computational simulations, Machine Learning (ML), and Natural Language Processing (NLP) to accelerate materials discovery. While automated quantum and classical simulations have established vast databases of material properties, recent advancements in NLP and Large Language Models (LLMs) now enable the massive, algorithmic extraction of materials data directly from scientific literature.

By integrating these diverse data sources to train robust ML algorithms, we can efficiently predict material behavior and bypass computationally expensive calculations. In this talk, I will present practical applications of these synergistic techniques within the energy sector. Specifically, I will discuss large-scale literature data extraction using NLP and LLMs for high-entropy oxide phase prediction. Furthermore, I will highlight the deployment of active machine learning workflows for selecting water/oil interface surfactants and optimizing nanoparticles for the Hydrogen Evolution Reaction (HER).