Gaussian processes and active learning: three recent chemistry-related developments
Event details
| Date | 18.03.2025 |
| Hour | 15:15 |
| Speaker | David Ginsbourger is heading the Uncertainty Quantification and Spatial Statistics Group and serving as Director of Studies in Statistics at the University of Bern, where I he is co-directing the Institute of Mathematical Statistics and Actuarial Science. At the University of Bern, he is also a member of the Oeschger Center for Climate Change Research, the Center for Artificial Intelligence in Medecine, and the Multidisciplinary Center for Infectious Diseases. On the editorial side, he is serving as Associate Editor of SIAM/ASA Journal on Uncertainty Quantification, Technometrics, and regularly as Area Chair / Meta-Reviewer for major Machine Learning conferences. |
| Location | Online |
| Category | Conferences - Seminars |
| Event Language | English |
In this talk, I will give an overview of three different research collaborations pertaining to Gaussian Process modelling and active learning algorithms applied to chemistry-related challenges. In particular, I will present recent work on targeted sequential design for uncovering high-activity molecules under precise and imprecise measurements schemes, and also present fast approaches to encode stochastic equivariances within kernel methods applied to dipole moment prediction.
Practical information
- General public
- Free
Organizer
- Andres M Bran, Rebecca Neeser, Jeff Guo, Philippe Schwaller
Contact
- Andres M Bran, Rebecca Neeser, Jeff Guo, Philippe Schwaller