Our laboratory has proposed and conducted automated experiments to evaluate the performance of cement dispersants. Through the abundant data obtained from these automated experiments, we can precisely capture the subtle responses influenced by the polymeric structure of cement dispersants.
The automated experimentation has the potential to transform the quality control paradigm of construction materials by integrating data science and machine learning. Before developing an automated experimental system for evaluating the performance of construction materials, our research team developed machine learning algorithms suitable for the characteristics of construction materials. Various algorithms have been developed, including observation-based learning and domain adaptation learning algorithms.
Kim JH, Kang IK, Shin TY, and Park CK (2025), Automated experimentation for evaluating cement dispersant performance, Cem. Concr. Res. 194: 107895, https://authors.elsevier.com/a/1kwJ%7E21ISwigq
Kang IK, Shin TY, and Kim JH (2025), Unbiased rheological properties determined by adversarial training with Bingham equation, Cem. Concr. Comp. 157: 105943, https://doi.org/10.1016/j.cemconcomp.2025.105943
Kang IK, Shin TY, and Kim JH (2023), Observation-informed modeling of artificial neural networks to predict flow and bleeding of cement-based materials, Constr. Build. Mater. 409: 133811, https://doi.org/10.1016/j.conbuildmat.2023.133811
Kim JH, Shin TY, Yekaterina S, and Park CK (2023), Data science approach to find an outlier in the group of cement dispersants, Constr. Build. Mater. 368: 130347, https://doi.org/10.1016/j.conbuildmat.2023.130347