(40d) New Materials Design Based on Machine Learning

Process-based chemical industries find actionable insights from data to design new products for improved performance and unique functionality. Development costs are lowered by combining the power of high-throughput data generation, information extraction and statistical analysis, proof of concept, fundamental understanding and empirical model development using machine learning. This report will cover three distinctive, yet connecting, elements of machine learning which enable materials design concepts. For example, process optimization with targeted application, pattern recognition in molecular structure design, and structure-property relationship in compressed representation together with case study examples.