Perspectives Towards AI and ML.

Abstract

This chapter discusses the applications of artificial intelligence (AI) machine learning (ML) to address the practical challenges in biomolecular crystallization processes. Although AI and ML techniques have been successfully applied in many scientific domains, to the best of our knowledge, there is no research that has yet applied ML to develop predictive models for biomolecular crystallization. Therefore, the key focus of this perspective is to demonstrate how ML approaches can be effectively applied to crystallization processes. While biomolecules and small molecules may have different properties and variables to predict, once sensor data are collected and transformed into a data array, the application of machine learning techniques becomes fundamentally similar from a mathematical point of view. Specifically, we start with a brief introduction to various types of ML methods and provide a review of the applications of AI and ML to crystallization. AI and ML have become increasingly important in improving our understanding of biomolecular crystallization, offering new insights and predictive capabilities that were previously unattainable. Subsequently, we introduce three ML approaches to address practical implementation issues in the context of modeling the crystallization processes. Two examples of crystallization processes are presented to demonstrate the effectiveness of the proposed ML algorithms.