Fred A. Hamprecht
Fred A. Hamprecht develops machine-learning algorithms for problems with spatial structure. He worked on image analysis for two decades before shifting toward quantum chemistry. His current focus is learning a kinetic-energy functional for density functional theory using geometric machine learning. In February 2026, Hamprecht’s Scientific AI group collaborated with a computational-chemistry group on STRUCTURES25, a neural method for stable orbital-free calculations of molecular energies and electron densities. His research connects mathematical methods, including combinatorial optimization and graph theory, with useful calculations in the natural sciences.
- Now at
- Heidelberg University
- Works
- 261
- Citations
- 15,951
- h-index
- 51
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
Earlier research
Learning from images
Hamprecht worked on image-analysis problems for roughly twenty years. Spatial structure is the continuing theme in his algorithm research.
-
As of October 2026
Geometric learning in chemistry
His work has shifted toward a long-standing quantum-chemistry problem: learning the kinetic-energy functional within density functional theory. His methods draw on geometric machine learning, combinatorial optimization and algebraic graph theory.
-
18 February 2026
STRUCTURES25
A neural-network method learns relationships between electron density and molecular energy. Hamprecht’s Scientific AI group collaborated with Andreas Dreuw’s computational-chemistry group. The tests concern organic molecules.
-
February–June 2026
Continued lab research
His research included function-centric graph-neural-network work in February 2026, and he received an ERC Advanced Grant in June 2026 for machine-learned orbital-free density functional theory. The work received continuing research support and used shared methods.
Sources and credits
Photo: ellis.eu; Fred A. Hamprecht profile photograph; creator not stated. License: Not stated by source. Source: ellis.eu. Cropped to a square and resized. Ask us to remove or replace this photo.