Senior Researcher
AI for Science achieved a breakthrough in the accuracy of DFT, a method for predicting the properties of molecules and materials, by using deep learning. This work can…
The Skala functional is a deep-learning based exchange-correlation functional trained to learn a truly non-local representation of the density to push the boundaries of accuracy in density functional…
The Microsoft Research Accurate Chemistry Collection (MSR-ACC) provides a collection of 150k+ accurate coupled cluster labels for training machine learning functionals. The MSR-ACC/TAE25 dataset releases 77k total atomization…
An overview of the development of density functional theory (DFT) over time, including the major milestones in its history and an outlook for the future of DFT.
My name is Sebastian Ehlert and I am a Senior Researcher in Microsoft Research AI for Science working pushing the boundaries for density functional theory (DFT) using deep learning and highly accurate and rigorous quantum chemistry methods at scale.
I am working on the Skala functional, the first step towards a truly data-driven development of exchange-correlation functionals to bring systematic improvability to DFT and provide accurate and robust models for all chemical space. Furthermore, I am driving the Microsoft Research Accurate Chemistry Collection (MSR-ACC) as our contribution towards the largest, most accurate and broadest dataset for training machine learning models with chemical accuracy (±1 kcal/mol).
Interested in testing out Skala? Please reach out to us in the DFT Research Early Access Program (opens in new tab).
I am also involved in the development of the extended tight binding (xTB) methods and an active open-source maintainer of many scientific libraries as well as a package maintainer on conda-forge to make the scientific software more accessible for everyone.
