Abstract

Machine learning and quantum chemistry are converging to make quantum simulations more powerful and accessible. In this talk, I will highlight how graph neural networks can predict excited-state densities, how learning from molecular orbitals boosts generalization and accuracy, and how our new library, LibRInt, enables automatic differentiation of quantum chemistry integrals. I will also introduce El Agente, our agentic AI framework for computational chemistry. I’ll conclude with a look at emerging directions at the intersection of machine learning and quantum simulation.

Biography

Dr. Abdulrahman Aldossary obtained his Ph.D. from the University of California, Berkeley working with Prof. Martin Head-Gordon. Dr. Aldossary worked on developing new computational quantum chemistry methods. These range from accelerating electronic structure calculations such as DFT to new methods for the analysis of these calculations to facilitate chemical interpretation. Dr. Aldossary obtained his HBS from Oregon State University working under Prof. Greg Herman. Since summer 2023, Dr. Aldossary has been conducting his postdoctoral work with Prof. Alan Aspuru-Guzik at the University of Toronto. In his free time, he likes to hike and bike, and explore new places.