Proceedings Paper • Open AccessSuperconducting Optoelectronic Networks for Quantum and Neuromorphic Computing
Efficient single-photon detectors have enabled new directions in experimental physics. Superconducting nanowire single-photon detectors (SNSPDs) are a vital component in loophole-free Bell inequality violation, metrology beyond the shot-noise limit, and nuclear clock transition sensing. We discuss recent progress in cryogenic silicon photonic platforms at the National Institute of Standards and Technology (NIST). These platforms co-integrate SNSPDs, silicon-on-insulator waveguides, superconducting amplifiers, and all-silicon light sources based on emissive crystal defect centers. Together, these elements lay a foundation for large-scale quantum information systems and extremely low-power neuromorphic architectures. Neuromorphic photonics uses the complementary properties of optics and electronics to realize information processing capabilities far beyond what is possible with pure electronics. Its performance potentials are highly desired as neural network approaches have retaken the helm of machine learning. We present recent results in neuromorphic photonic architectures at NIST and Princeton: one signaling at single-photon levels, one signaling on sub-nanosecond timescales, both propelled by the emerging industrialization of silicon photonics. Neuromorphic and quantum processing approaches based on silicon photonics will overlap substantially in terms of technical challenges and enabling technologies.