SEMINAR

Biomaterials and Tissue Engineering Research Center

Shanghai Institute of Ceramics, Chinese Academy of Sciences

中国科学院上海硅酸盐研究所生物材料与组织工程研究中心

Machine Learning and AI-Driven Design of Calcium Carbonate-Based Bioceramics for Biomedical Applications

Speaker：Assoc. Prof. Bogdan Parakhonskiy

(Ghent University, Belgium)

报告时间：2026年04月09日(星期四) 10:00-12:00

报告地点：长宁园区4号楼14楼第一会议室

联系人：朱钰方 研究员（52411808）

Assoc. Prof. Roman Chernozem

Personal information:

Bogdan Parakhonskiy is a professor (PI) at Ghent University, Belgium, in the NanoBioTechnology Group, Department of Biotechnology, Faculty of Bioscience Engineering. He received his PhD in Physics from Lomonosov Moscow State University in 2009. From 2010 to 2014 he held a Marie Curie Fellowship at the University of Trento, Italy, and later led research activities in theranostic nanomaterials at Saratov State University. Since 2015 he has worked at Ghent University with support from the Research Foundation - Flanders (FWO).

His research focuses on bioceramic and hybrid micro- and nanomaterials, especially calcium carbonate-based systems for loading, release, sensing, and regenerative medicine. In recent years he has integrated machine learning, explainable AI, SHAP-based analysis, and active learning into materials synthesis optimization, crystallization control, and image-based analysis in materials science.  have authored 132 peer-reviewed publications and has an h-index of 42. My work has appeared in high-impact journals including Adv. Colloid Interface Sci. (IF 19.3), Nano Energy (IF 17.1), Angew. Chem. Int. Ed. (IF 16.9). 

ABSTRACT:

Porous bioceramic materials are of broad interest for biomedical applications because they can combine structural tunability, degradability, cargo loading, and biological functionality within a single platform. However, the development of such systems is still often driven by slow, empirical experimental screening, which makes it difficult to efficiently optimize their composition, morphology, porosity, and performance. In this talk, I show how machine learning and data science can accelerate the design of calcium carbonate-based bioceramics and establish clear links between synthesis conditions, material structure, and functional behavior.

Our work focuses on porous calcium carbonate micro- and nanoparticles, hybrid mineral-polymer composites, and multifunctional CaCO₃ systems with tunable phase composition, morphology, porosity, degradability, and loading capacity. By combining experiments with machine learning models, SHAP-based explainable AI, and active learning, we identify the key variables controlling crystal phase formation, particle size, and structural properties, while also selecting the most informative next experiments for efficient optimization.

These data-driven strategies enable the rational development of new mineral and hybrid bioceramic materials with controlled properties and multifunctional performance. The presentation will show how machine learning, SHAP-based explainable AI, and active learning guide the design of calcium carbonate systems, together with representative applications in drug delivery, plant-related delivery, sensing, regenerative medicine, and bone reconstruction.