(Bio)Chemical Process Development aided by Robotics and Machine Learning

In this talk I’ll describe the concept of a fully digital (bio)chemical R&D platform and how access to new tools is rapidly transforming the nature of R&D in molecular sciences. I’ll discuss concepts of molecular and reaction parameterisation that enable transfer of molecular R&D into the domain of mathematics, sometime avoiding, and sometime enhancing the classical physics-based computational methods. Specifically, I’ll present the concept of representing reaction data as a graph and using network optimisation for developing synthetic strategies. Once a route has been identified, individual reaction predictions and optimisation can be facilitated by ML algorithms and chemical robotic experiments. Finally, time permitting, I’ll introduce the application of reinforcement learning to identification of optimal process options.