The Potential of Machine Learning to Strengthen the Microwave-Microfluidic Characterization of Biological and Chemical Materials

Dielectric spectroscopy allows for label-free, non-invasive, and rapid characterization of biological and chemical materials. Its integration with microfluidics opens the door to measurements on low sample volumes. The acquired measurement data are frequency-dependent, complex vectors, containing ample information on the material under test (MUT) at multiple scales. Extensive post-processing algorithms and interpretation methods can give meaning to the results. In recent years, machine-learning algorithms have been applied to facilitate this information extraction. This talk discusses the opportunities of machine learning to strengthen the dielectric spectroscopy analysis and uncover detailed insights into the biological and chemical properties of the MUTs.