A 5x performance increase for formulations, chemicals, and materials
This white paper discusses how an enhancement to the Alchemite™ methodology offers greater accuracy and faster solutions when targeting specific goals in R&D projects.
Executive Summary
Machine learning shows great potential to design and understand formulations, chemicals, materials, and biopharmaceuticals. Being focused on the end goal is key: to find the mixture that will fulfil a target specification, for example, to maximize strength, minimize cost, or minimize carbon footprint. We demonstrate how Alchemite™ offers a machine learning workflow that uniquely exploits the target specification to train a machine learning model that is optimized for this specific goal. This allows a 5x increase in performance over a model trained for accuracy over all formulations, saving time and money in the design process.
