Machine Learning for Chemical Synthesis Planning and Development

Abstract

Data Science has become increasingly valued as a tool for accelerating chemical discovery.  It holds great promise to improve the efficiency of designing synthesis of molecules. Leveraging machine learning algorithms and rich historical chemical reaction data, efficient solutions can be developed to undertake tasks that previously require heavy involvement of expert chemists. Specifically in this talk, I will discuss about the use of machine learning to recommend reaction conditions, including catalysts, solvents, reagents and temperature, for organic synthesis development. I will also demonstrate how we can leverage this information to optimize synthesis plans for or molecule libraries, with the goal of encouraging shared pathways and reducing chemical inventory.

Biography

Hanyu Gao is an assistant professor at the Hong Kong University of Science and Technology. Hanyu obtained his Bachelor’s degree in chemical engineering from Tsinghua University in 2012, where he was Magna Cum Laude. He then completed his Ph. D. in the Department of Chemical and Biological Engineering at Northwestern University with Prof. Linda Broadbelt. After that he worked as a postdoctoral associate at MIT in Prof. Klavs Jensen’s group from 2017 to 2020. Hanyu’s research interest lies in using modeling techniques, including simulation, optimization and machine learning, to solve chemical engineering problems ranging from polymer reaction engineering to organic synthesis design.