Acceleration of Toothpaste Development using Bayesian Optimization

Abstract

With the diversification and sophistication of consumer needs, it is important to increase the productivity of each researcher in the field of research and development. In general, toothpastes are composed of multiple ingredients such as cleaning agents, wetting agents, foaming agents, binders, and medicinal ingredients. In addition to the experience and knowledge of researchers, a certain period of time is required to develop high-quality products that meet consumer needs and there is an urgent need to improve the accuracy of research and speed up the development process. In this study, a Bayesian optimization method, a type of machine learning, was employed to shorten the development period of toothpaste. As a result of repeated experiments with the optimized composition based on the composition and physical property information, the toothpaste composition satisfying the target physical property values was secured in half the expected study period of the conventional method.