Emerging AI-enhanced approaches for polymer design

The ability to generate large data sets from experiments and simulations, coupled to the development of new machine learning algorithms, is leading to significant changes in polymer physics reasearch. In this work, I will discuss these advances in the context of two examples. In the first, large data sets from simuations are used to train neural networks for prediction of polymer properties as a function of monomer sequence and chemical characteristics. In the second , large data sets from simulations are combined with machine learning to predict the long-time relaxation of microphase separated polymeric materials.