Photonic In-Memory Computing Primitive for Spiking Neural Networks Using Phase-Change Materials

Spiking neural networks (SNNs) offer an event-driven and more-biologically-realistic alternative to standard artificial neural networks based on analog information processing. This can potentially enable energy-efficient hardware implementations of neuromorphic systems that emulate the functional units of the brain; namely, neurons and synapses. Recent demonstrations of ultrafast photonic computing devices based on phase-change materials (PCMs) show promise for addressing limitations of electrically driven neuromorphic systems. However, scaling these stand-alone computing devices to a parallel in-memory computing primitive is a challenge. In this work, we use the optical properties of the PCM Ge2Sb2Te5 to propose a photonic SNN computing primitive, comprising a nonvolatile synaptic array integrated seamlessly with previously explored integrate-and-fire neurons. The proposed design realizes an in-memory computing platform that leverages the inherent parallelism of wavelength-division multiplexing. We show that the proposed computing platform can be used to emulate a SNN inferencing engine for image-classification tasks. The proposed design not only bridges the gap between isolated computing devices and parallel large-scale implementation but also paves the way for ultrafast computing and localized on-chip learning.