Abstract: How can we best take advantage of the growing knowledge and tools in artificial intelligence (AI) and machine learning (ML) to design and operate processing and biomanufacturing systems optimally and advance the fundamental understanding of the underlying phenomena for these systems? Unlike many of the native application domains of AI/ML algorithms, these processes generally yield relatively structured data sets that mask information and are not abundant. Also, there is significant accumulated know-how regarding these systems. Therefore, employing AI/ML techniques for these systems requires selecting the appropriate method and customization to incorporate the existing knowledge.
Hybrid modeling combines first-principles models and data-driven models, which can be based on AI/ML algorithms. Within the hybrid model, the first principles model implements known process knowledge, and the data-driven model compensates for the disagreement caused by an incomplete understanding of the process mechanism and enhances the accuracy of the model. This talk will highlight the strengths and challenges of building hybrid models for engineering applications using three seemingly disparate examples: (1) plant optimization for minimizing energy consumption, (2) accurate predictions of liquid entrainment fraction and its uncertainty, and (3) modeling differentiation of human-induced pluripotent stem cells to cardiac muscle cells.

Speaker Bio: Dr. Cremaschi is the Chair, B. Redd & Susan W. Redd Endowed Eminent Scholar Chair Professor, and Head of the Cremaschi group in the Department of Chemical Engineering, Auburn University. Her research focuses on risk management, optimization, process synthesis, machine learning, and planning under uncertainty. Her research group develops systems analysis and decision support tools for complex systems, mainly focusing on the biomanufacturing, pharmaceutical, and energy industries. She is a recipient of the NSF CAREER award (2011), the Zelimir Schmidt Award for Outstanding Researcher (2013), and the Senior Research Award for Excellence (2021), among others. She is a member of the 2018 Class of Influential Researchers selected by the Industrial and Engineering Chemistry Research journal. Her research work has been consistently supported by industrial collaborations in addition to federal agencies. She also serves on the Dow Chemical Company Digital Technology Advisory Board. She earned a Ph.D. from Purdue University and an M.S. and B.S. from Bogazici University (Turkey), all in chemical engineering.