Applications of Artificial Intelligence Models for Computational Flow Dynamics and Droplet Microfluidics

Abstract

Microfluidics allows for the manipulation and analysis of minuscule amounts of liquid within a system that contains multiple channels, ports, and samples. Advanced microfluidic technology can incorporate numerous functional units onto a tiny chip made of glass, plastic, or polymers. By combining microfluidic systems with artificial intelligence (AI) models, it is possible to optimize the design and testing processes, leading to increased automation and intelligence in experiments. The AI models are divided into four general categories: unsupervised learning, supervised learning, semi-supervised learning, and reinforcement learning. These models prove invaluable in discovering and optimizing chemical synthesis, which can be costly and time-consuming. Additionally, AI models aid in simulating the assembly of colloidal materials in microfluidics, speeding up the prediction of material characteristics necessary for designing novel materials with interesting physical or chemical properties. Similarly, AI algorithms can predict the behavior of multiphase fluids, assisting in the design of microfluidic chips for various applications. Another application of AI models in microfluidics involves the detection of cellular matter, including DNA, RNA, proteins, and other metabolites, using droplet-based biotechnology techniques. Machine learning techniques can be employed to segment and classify droplets in images for this purpose. In this review, we cover the known applications of AI algorithms in the design of microfluidic systems and flow techniques, including droplet microfluidics. In this review, we cover the known applications of AI algorithms in the design of microfluidic systems and flow techniques, including droplet microfluidics.