Riassunto analitico
Recent advancements in nanoelectronics technology have led to the development of cutting-edge massively parallel nano-biosensing platforms, which promise to create a new paradigm for personalized medicine. One of the key features of
nanoelectronics-based biosensors is their ability to operate with low sample volumes, resulting in portable low-cost devices easily interfaced to the outside world.
The use of nanomaterials and nano-electrodes in biosensors has also led to improved sensitivity and selectivity, which results in a remarkable improvement in the ability to resolve negligible concentrations of analytes in biological fluids.
However, one of the key challenges in biosensing is the accurate detection and analysis of small analytes (e.g. micro and nano-particles or small biomolecules), which can be difficult due to the complexity of biological systems.
Machine learning has emerged as a promising tool for improving the detection performance in many contexts, but up to now, very limited work has been carried out on nano-biosensors.
In this work, we investigate the potential advantages of utilizing machine learning algorithms, particularly neural networks, trained on simulated data, to accurately predict the size of dielectric micro-particles measured by a CMOS nano-capacitor biosensor in two different solutions. The use of simulated data for training enables the generation of large and diverse datasets with known micro-particle sizes, which
can help in training the neural network models and validating their accuracy. The results allowed us to identify methods to generate training datasets by means of physical CV-FEM simulations of the nano-capacitor array; and led us to identify
suitable neural network architectures to extract analyte physical parameters such as particle radius. They also indicate that training on simulated datasets empowers the machine learning algorithms and improves the sensor's ability to accurately predict the particle size. These promising results pave the way to extend the use of the nano-capacitor platform to relevant applications for biology and environmental studies.
|
