Physical Unclonable Function in Spiking Neural Network Based on in-Te Ovonic Threshold Switching Devices

This work presents a threshold-adjustable leaky integrate-and-fire (LIF) neuron based on <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{In}_{3} \text{Te}_{7}$</tex> ovonic threshold switching (OTS) devices with high endurance and low leakage current. A two-layer spiking neural network (SNN) built with the neurons achieves 85 % accuracy in handwritten digit recognition. Additionally, a physical unclonable function (PUF) based on the LIF neurons demonstrates high uniqueness and reliability. This hardware solution integrated with neuromorphic computing system offers a new approach for data security, which is crucial for AIoT.