Assist.-Prof. Mike Preuss is Assistant Professor of Artificial Intelligence at the Leiden Institute of Advanced Computer Science (LIACS). Though originally from the field of evolutionary optimization, he has over 10 years of experience in machine learning/game AI and several successful collaborations with chemistry and engineering researchers. He is a renowned scientist in the fields of multi-modal and multi-objective optimization, search and reinforcement learning algorithms, and machine learning methods that are first explored in the area of computer games but then transferred to real-world problems. His algorithms show great potential for solving real chemical synthetic tasks where all reagents, reactants, catalyst, solvent, and temperature needs to be taken into account to optimise the outcome and the yield of reactions and reaction planning. The Fellow will learn and implement the algorithms to select appropriate conditions for the reactions and also score alternative retrosynthetic pathways suggested by the models. Capabilities in multi-objective optimization in this area are critical for the success of the AiChemist project.
Aske Plaat, PhD, Scientific Director of the institute and head of the reinforcement learning group, will co-supervise the project of DC5, in addition to conceptualising and assessing their research.