Deep learning for DFT

Опубликовано 18 июня 2025, 15:39

In Density Functional Theory, the exchange correlation functional captures the complex relationship between its input—the electron density — and its output: the exchange-correlation energy. The electron density is represented as a large, irregular point cloud, where each point encodes local information such as the density itself and its gradient. An accurate functional must capture intricate non-local interactions across the electron density, at manageable computational cost.

Read more about DFT: microsoft.com/en-us/research/b...
Microsoft Research Accurate Chemistry Collection: ai.azure.com/labs/projects/msr...