Polymer is an important class of materials that are used in almost every aspect of our daily lives. The property of polymers has great tunability due to the infinite possibility of polymer microstructures, including the variation of monomer types, polymer chain length, composition and sequence distributions. On the other hand, this also creates a large design space for polymers and it is becoming increasingly important to develop computational approaches to simulate polymerization systems to design the reaction recipe.

It is therefore necessary to navigate the design space effectively. Active machine learning algorithms have been developed for this purpose, where a machine learning is iteratively trained, with a strategy to maximize the information gain with new data collected. In this project, you will use kinetic Monte Carlo (KMC) simulations for generating polymerization reaction data, and an active learning framework to maximize the efficiency of (simulated) data collection.