The article discusses the development of neuromorphic computing systems that combine superconducting electronics and integrated photonics to create artificial neural networks. The goal is to build systems that can process information in a way similar to the human brain, enabling more powerful artificial intelligence.
Key developments include:
1. Synaptic circuits with integrated local memory that can be programmed to define synaptic weights
2. Superconducting pixels for single-photon sensitive cameras
3. Superconducting optoelectronic synapses that perform weighting and leaky integration of single-photon signals
4. Ultrahigh impedance superconducting thermal switches for interfacing superconductors with semiconductors and optoelectronics
5. Multi-planar integrated photonic routing manifolds for neural networks
The research team has developed a simulation framework that increases simulation speeds by a factor of ten thousand while maintaining close agreement with the brute force approach. The software is publicly available.
The ultimate goal is to pursue optoelectronic intelligence by integrating optical components for communication with superconducting electronics for computation to enable large-scale neural systems. The research team is currently seeking postdoctoral researchers to join their efforts.
Source: https://www.nist.gov/programs-projects/physics-and-hardware-intelligence
Keywords: Superconducting, Optoelectronic, Neuromorphic, Photonic, Josephson
