Brain‐Like Computing Using CuInS ₂ Quantum Dots: A Synergy of Materials, Devices, and Machine Learning Modeling

ABSTRACT The advanced artificial intelligence applications demand new computing paradigms that can mimic the brain's processing capabilities. Chalcogenide‐based quantum dots (QDs) possess excellent chemical stability and electrical properties, emerging as promising candidates for neuromorphic devices. In this work, we have developed CuInS 2 QDs (CIS QDs) based resistive switching (RS) devices for non‐volatile memory and neuromorphic computing applications. The CIS QDs‐based RS device can switch at +1.15 V (V SET ) and −0.73 V (V RESET ) and demonstrate good non‐volatile memory properties (endurance: 5 × 10 3 cycles and retention: 10 4 s). The switching variability of the device was modeled, predicted, and forecasted by using multivariate time series and machine learning techniques. Additionally, the CIS QD‐based RS device has mimicked various forms of synaptic learning functions such as potentiation, depression, excitatory post‐synaptic current, and inhibitory post‐synaptic current. Moreover, the multilayer perceptron neural network was implemented to demonstrate the handwritten digits classification ability of the CIS QDs‐based RS device. These results demonstrate the potential of CIS QDs for next‐generation neuromorphic computing hardware.