Advances in synthesis methodology can have an impact on the selection of starting materials, synthesis process steps, and isolation of energetic materials. Current industrial processes for the production of energetic materials do not take into account hazardous materials and wastes. The application of modern and novel computational tools can reduce the environmental footprint of these large scale industrial processes. This short course covered recent research and development of machine learning techniques for application to energetic materials. Instructors covered two distinctly different topics for model development: property prediction and synthetic chemistry planning. They reviewed the state-of-the-art, describe some recent DoD results, and provide practical advice for getting started in the field. Data collection and curation, data representation, classes of algorithms (e.g., neural network models), and model evaluation methods also will be discussed.


Dr. Nirupam Trivedi, DEVCOM ARL

Machine Learning for Molecules

Dr. Brian Barnes, DEVCOM ARL

Machine Learning Models for Synthesis Planning

Dr. Igor Schweigert, Naval Research Laboratory (NRL)