(368bo) A Hybrid Approach for Modeling CHO Cell Culture That Incorporates Explainable Machine Learning

The production of monoclonal antibodies (mAbs) using Chinese Hamster Ovary (CHO) cell cultures is fundamental to biopharmaceutical manufacturing. However, challenges remain in building predictive models of cell culture processes due to the complexity of biological systems^[1]. While machine learning approaches can be effective for building model for processes with poorly understood phenomena, building models that produce highly accurate predictions typically requires large amounts of data collecting over a wide range of operating conditions, which is time-consuming and expensive to obtain.
This study presents a hybrid model for CHO cell culture that integrates domain knowledge into a machine learning model. In this structure, the machine learning model estimates the kinetic parameters of a mechanistic model, which uses machine learning to describe the effects of phenomena for which mechanistic models are not yet established^[2]. A sparse machine learning method is employed to enhance model interpretation. This hybrid model structure significantly reduces data requirements and increases reliability while incorporating explainable machine learning to help develop an understanding of relationships between process variables for phenomena in which mechanisms are not yet well established.

[1] M. S. Hong, K. A. Severson, M. Jiang, A. E. Lu, J. C. Love, and R. D. Braatz. Challenges and opportunities of biopharmaceutical manufacturing control. Computers & Chemical Engineering, 110:106-114, 2018.

[2] M. Aykol, C. B. Gopal, A. Anapolsky, P. K. Herring, B. van Vlijmen, M. D. Berliner, M. Z. Bazant, R. D. Braatz, W. C. Chueh, and B. D. Storey. Perspective—Combining physics and machine learning to predict battery lifetime. Journal of The Electrochemical Society, 168(3):030525, 2021.

Research Interests

My research lies at the intersection of Artificial Intelligence/Machine Learning (AI/ML) and process systems engineering. I focus on using AI/ML technologies in process modeling and optimization to enhance system efficiency and improve decision-making processes. Specifically, I am interested in developing hybrid models that integrate conventional engineering approaches with advanced AI/ML technologies, aiming to improve predictive accuracy, reduce operational costs, and maximize the performance of complex systems.