The Institute of Biological Chemistry cordially invites you the following lecutre as part of the Institute's Seminar Series:

Ass.-Prof. Dr. Esther HEID, Faculty of Technical Chemistry, TU Wien

"Predictive and Generative AI for Chemistry"

Abstract: Machine learning is increasingly used to predict molecular properties, analyze chemical reactions, and explore new regions of chemical space. These methods learn relationships directly from chemical data and can support tasks such as estimating physico-chemical properties, reactivity, or biological activity. In practice, however, chemical datasets are often small, costly to obtain, and biased, making the development of data-efficient models a key focus in the field.
Successful predictions typically rely on architectures that incorporate domain knowledge, such as the chemical and physical symmetries of a system. While such domain knowledge can be added easily to molecular property prediction, the meaningful description of chemical reactions, i.e. the change of a set of molecule to a set of products is much more challenging. In this talk, I will provide an overview of AI and machine learning for chemical applications, and present my work on modeling molecular and reaction properties, which a focus on developing new reaction representations and architectures. Finally, I will discuss advances toward utilizing generative models such as diffusion and flow matching for reaction pathway prediction.