Droplet based microfluidics integrated with machine learning

Machine learning for microfluidic design and control

“…Droplet microfluidics is an application well-suited for MLbased performance prediction, as the complex fluid dynamics of multiphase flows prevent generalizable understanding. 31,79 In Khor et al, ML models predicted droplet stability within tightly packed emulsions passing through a constriction. 77 The developed model, an 8-dimensional convolutional autoencoder for feature extraction and a two-layer fully connected classifier, was trained on 500 000 droplets and could predict droplet stability with 91.7% accuracy, in contrast to the 60% accuracy of conventional scalar descriptors (Fig.…”

Intelligent control of nanoparticle synthesis through machine learning

“…101 Before building the machine learning model, we can normalize the data of the prepared dataset, which can improve the accuracy and computational convergence speed of the machine learning model. 102,103…”

Numerical Investigation of the Influence of Microchannel Geometry on the Droplet Generation Process

“…The development of machine learning (ML) capabilities have also been explored to try and overcome challenges in microfluidic technology and biomedical and biotechnology applications. The integration of both Droplet Based Microfluidic (DBMF) and ML facilitates the development of highprecision, automated and optimal tools for multiple applications (Srikanth et al 2021). Other aspects of microdroplet interaction occur in proton exchange membrane fuel cells (PEMFCs); their channels see a considerable impact on the wall, and this affects cell performance (Z.…”