CAPACITIVE ARTIFICIAL NEURAL NETWORKS
Dr. Qiangfei Xia
Applications include pattern recognition, image processing, natural language understanding, robotics, and other fields where machine learning and artificial intelligence techniques are applied.
The invention introduces a technology that utilizes capacitative components to implement artificial neural networks, enabling advanced computing systems with improved efficiency and performance.
It addresses the need for more efficient and scalable artificial neural network architectures. It aims to provide a solution that allows for faster and more energy-efficient information processing, enabling the development of advanced machine learning and artificial intelligence systems.
The patent describes a novel approach that incorporates capacitative components into the design of artificial neural networks. Capacitors are utilized as key elements in the network's structure, taking advantage of their ability to store and release electrical charge.
The capacitative artificial neural networks (CANNs) outlined in the patent leverage the inherent properties of capacitors to facilitate fast and parallel processing of information. The capacitors store electrical charges that represent synaptic weights, enabling the network to perform computations efficiently.
The invention further describes the integration of CANNs with other components such as transistors, amplifiers, and interconnects to create complete neural network systems. These systems demonstrate improved performance in terms of speed, energy efficiency, and scalability compared to traditional artificial neural network architectures.
Available for Licensing and/or Sponsored Research
UMA 18-003
US Patent Issued 10,740,672
|
