전략적인 머신러닝 기반 설계방법 및 제조업 적용사례 - 유승화 교수(KAIST 기계공학과)

In this Merric Webinar presentation, we will explore how research and development trends in materials science and the manufacturing industry have been significantly transformed by the emergence of machine learning (ML). Data-driven methodologies have accelerated the discovery, production, and product development of new materials by accurately and swiftly predicting complex physical processes and mechanisms that cannot be fully explained by traditional material or physical theories through correlations in data. However, despite the availability of various ML models, the most important and challenging question remains: which ML algorithms are best suited for the design problems we face? In this presentation, we aim to segment challenges in material/shape/process design into categories based on data characteristics and design space size, providing guidelines for selecting ML algorithms to address these questions.