A centralized random hopping neural network circuit structure and a design method thereof

Abstract

The invention discloses a centralized random hopping neural network circuit structure and a design method thereof. The circuit structure comprises a neural network module, a random number generation module and the process control module. Whether to start a random hopping process or not is decided according to the output of the process control module, and when the random hopping process is started,the operation of the random number generation module is controlled at the same time, namely two sub-modules of the random number generation module are controlled, and the two random number generatorsare controlled to output two groups of random numbers; the neural network module is used for realizing a neural network and processing the training set; the random number generation module comprisestwo independent random number generators and is used for generating two groups of random number groups and realizing weight change at a random node of the neural network; and the process control module is used for controlling the change of the working mode of the whole circuit structure. Centralized processing is adopted, the weight and the node address of the neural network are more convenientlychanged, and automatic processing is achieved.

Description

technical field [0001] The invention relates to an algorithm and a design method of a neural network chip, and belongs to the field of digital integrated circuit design. More specifically, the invention relates to a method for applying a neural network algorithm more deeply. Background technique [0002] Artificial Neural Network (ANN) is a research hotspot in the field of artificial intelligence since the 1980s. It is an extensive parallel interconnected network composed of adaptable simple units. His organization can simulate biological nerves. The system's interaction with real-world objects. It abstracts the human brain neuron network from the perspective of information processing, establishes a simple model, and forms different networks according to different connection methods. In engineering and academia, it is often referred to directly as a neural network or a neural network. A neural network is an operational model consisting of a large number of nodes (or neuron...

AI Technical Summary

Problems solved by technology

[0005] Today's neural networks perform the assigned tasks in a stable state, but the real brain does not always produce results in a stable state, often the unstable state brought by the outside world will lead to the generation of new inspiration

Images