Volatile tin oxide memristor for neuromorphic computing

Abstract

The rise of neuromorphic systems has addressed the shortcomings of current computing architectures, especially regarding energy efficiency and scalability. These systems use cutting-edge technologies such as Pt/SnOx/TiN x /TiN memristors, which efficiently mimic synaptic behavior and provide potential solutions to modern computing challenges. Moreover, their unipolar resistive switching ability enables precise modulation of the synaptic weights, facilitating energy-efficient parallel processing that is similar to biological synapses. Additionally, memristors' spike-rate-dependent plasticity enhances the adaptability of neural circuits, offering promising applications in intelligent computing. Integrating memristors into edge computing architectures further highlights their importance in tackling the security and efficiency issues associated with conventional cloud computing models.