Researchers at NIST have developed a neuromimetic circuit that mimics the structure and function of biological neurons using optoelectronic neural networks. The system consists of interconnected processing units (neurons) connected by integrated photonic waveguides, where signal strength and connectivity can be dynamically adjusted.
The key components include:
– Superconducting photon detectors and electronics
– Waveguide-integrated few-photon LEDs
– Reconfigurable nanophotonic waveguides
This optoelectronic platform achieves event rates of 20 MHz, 100x faster than recent CMOS systems, with zero power consumption in steady state. The use of optical signals allows for encoding information in multiple degrees of freedom, including photon number, frequency, and pulse rate, enabling complex neuromorphic computing.
The technology has the potential to achieve states of complexity rivaling biological systems while maintaining ultra-low power densities, making it promising for future neuromorphic computing applications.
Source: https://www.nist.gov/patents/neuromimetic-circuit
Keywords: optoelectronic, synapse, photons, superconducting, neuromorphic
