I will begin with a general overview of modern density functional theory and its applications, from room temperature superconductors to how well fish can smell. I will explain what DFT is, why it's so useful, and its current limitations. This should be accessible to anyone who has a general knowledge of quantum mechanics. In the second half, I will describe efforts to use modern techniques of machine learning, including neural networks and differentiable programming, to create new density functional approximations that are nothing like those that people create. I will report on some of the most recent work I've been involved with, namely finding approximations that work for strong correlations.
About the speaker:
Kieron Burke is a Chancellor's Professor of UC Irvine. He is also a fellow of the American Physical Society, the British Royal Society for Chemistry, and the American Association for the Advancement of Science, and a member of the International Academy of Quantum Molecular Sciences. He is known around the world for his many educational and outreach activities. According to google scholar, his research papers are now cited more than 21,000 times each year. Prof. Burke’s research focusses on developing all aspects of density functional theory: formalism, extensions to new areas, new approximations, and simplifications. His work is heavily used in materials science, chemistry, matter under extreme conditions (such as planetary interiors or fusion reactors), magnetic materials, molecular electronics, and so on. He has given talks in theoretical chemistry, condensed matter physics, applied mathematics, computer science, and even organic chemistry.