A Machine Learning Approach to Catalysis Modeling
Date:
Abstract: Tuning machinery for optimal output is a major step in catalysis setup. Each iteration costs practitioners money and time even with thorough understanding of the physical processes involved. With standard modeling techniques, computational costs suffer from the same downfall. To alleviate this step in the process, we deal with discrete approximations of filters and train neural networks to generate similar structures to estimate tuning parameters. We will discuss how we use Graph VAEs to generate bounds for tuning parameters as well as the statistical foundation of this technique.