AI Tools for Material Enhancement and Discovery in Energy Applications by Prof Eliodoro Chiavazzo

The accelerated development of advanced materials is key to transformative progress in energy applications. In this talk, I will present recent advancements from my group at the intersection of artificial intelligence (AI), energy, and materials science, focusing on AI-driven strategies for material performance enhancement, novel material discovery, and result generalisation using complex data-driven models.

First, I will discuss active learning strategies for enhancing material performance. Traditional optimisation is time- and resource-intensive. By integrating active learning with modeling workflows, we can select the most informative experiments to accelerate improvements. I will showcase examples from our recent work on thermal energy storage and solar fuels. Next, I will address new material discovery. The goal is to predict promising materials yet to be synthesised. I will present our latest work combining machine learning, domain expertise, and generative modeling, and address cross-domain data bias and ways to mitigate it.

Prof Eliodoro Chiavazzo is Full Professor of Applied Physics at the Department of Energy, Politecnico di Torino. He serves on the PhD Faculty Board in Energetics and was formerly Deputy Coordinator of the Mechanical Engineering program. He earned his PhD from ETH Zurich and was a Fulbright Fellow at Princeton University in 2013. As Director of the Multi-Scale Modeling Laboratory (SMaLL), he leads 25+ researchers on atomistic to macroscopic modeling for energy storage, solar desalination, and cooling systems. He leads large courses in thermodynamics and energy systems, and reviews for ERC, U.S. DOE, DFG, and Italy’s “Rita Levi Montalcini” program. He sits on the editorial boards of Scientific Reports (Springer Nature) and Entropy (MDPI), and has led EU-funded projects while collaborating with MIT and other global institutions.