Platonic Representation of Foundation ML Interatomic Potentials I LeMaterial Reading Group

Paper Link: https://www.nature.com/articles/s4225... Foundation machine learning interatomic potentials (MLIPs) have emerged as powerful tools for atomistic simulation, yet different models encode chemical environments in incompatible latent spaces, limiting direct comparison and interoperability. The platonic representation hypothesis suggests that sufficiently capable models converge towards a shared statistical representation of reality. Will this hold for MLIPs and what materials insights we can get from their representations? In this talk, Dr. Zhenzhu Li will share their recent story trying to answer these questions. Dr. Zhenzhu Li is an AI in Science Fellow at Imperial College London. Her research lies at the intersection of computational materials science and artificial intelligence, aiming to address the urgent need for AI-driven materials discovery by elucidating defect–structure–property relationships and advancing intelligent generative frameworks for next-generation photovoltaic materials. LeMaterial Reading Group is a recurring gathering where we discuss recent papers at the intersection of AI, chemistry and materials science.