Professor Paffenroth Receives a 3-Year Grant to Develop Data Science Driven Approaches for Chemical Identification

Real-time chemical identification and classification are crucial for many reasons and accurate identification enables accurate assessment of health and safety risks, as well as effective planning and preventive measures. Machine learning classification models serve as a valuable tool to this end. However, these models typically require extensive datasets to effectively generalize beyond their training data. Due to the high time and financial costs associated with generating chemical datasets at scale, producing sufficient data to train chemical classification models poses a significant challenge. Prof. Paffenroth, Core Data Science Faculty Member and Associate Professor of Mathematical Sciences, and his Data Science PhD student Cate Dunham recently received a $1,000,000 over 3-year project with the Defense Threat Reduction Agency and the US Army DEVCOM-SC to develop state-of-the-art methods to develop deep neural networks for the generation of synthetic chemical signatures.