data-driven system design
ML Algorithms, Autonomous Experimentation, High-throughput Simulation/Experiment
The properties of a given material are typically influenced by a host of parameters, which makes it not easy to determine the properties and find desired, optimal properties. This is true for polymers as well. Key chemical information encoded in building blocks affects intra- and intermolecular interactions, which ultimately tune nano- and microscale structures and macroscale behavior. The associated parameter space is vast and, therefore, exploring the space is not trivial.
We take full advantage of every single simulation and/or experiment by deploying autonomous experimentation (AE) methods, in which machine learning (ML) algorithms enable autonomous data acquisition with surrogate model computation and uncertainty quantification in the associated parameter space and thereby high-throughput simulations/experiments.