Use of machine learning/artificial intelligence in chemical sensors and biosensors

This is an introductory chapter for the subsequent chapters (Chapters 4–16). It summarizes how machine learning and artificial intelligence have been used for various chemical and biological sensing applications. Examples include high-dimensional data that enables the detection of a wide array of chemicals and biomolecules without needing a specific receptor. They are categorized by the increasing order of data complexity: sensor arrays (electronic nose), spectra (Fourier transform infrared, Raman, and surface-enhanced Raman spectroscopy), and images (microscopic, proximal, and hyperspectral). Unconventional data follow it, including a set of biological polymers and the capillary flow velocities on microfluidic chips. Finally, an extensive array of conventional chemical sensor and biosensor data is introduced, which can be used to monitor environmental contaminants in the ecosystem, to study complex protein dynamics at solid–liquid interfaces, and to enable single-molecule analysis for disease diagnosis and personalized therapeutics.