Harnessing Cross-Disciplinary Data for R&D Efficiency

This technical report discusses Lion Corporation's innovative approach to enhancing research and development (R&D) productivity through the application of proprietary transfer learning models. The report outlines how the R&D Division is utilizing materials informatics to accelerate value creation, specifically by developing a machine learning model that predicts the quality of body washes with new ingredients. This model has been adapted to apply research data from body washes to the development of bath detergents, enabling the use of over 100 years of accumulated research data. The report details the challenges faced in the early stages of product development, particularly the limitations of data volume, and how the new methodology addresses these issues. The findings were presented at the 48th Chemoinformatics Symposium, where the research received the Outstanding Poster Award for the second consecutive year. The report concludes with the potential for this methodology to significantly shorten R&D cycles and improve overall productivity in product development.