Artificial neural network construction method and system inspired by a brain inhibition feedback mechanism

Abstract

The invention provides an artificial neural network construction method and system inspired by a brain inhibition feedback mechanism. The construction method comprises the following steps: distinguishing neurons in an artificial neural network of a plurality of reference samples into excitative neurons and inhibitory neurons; inhibiting the inhibitory neurons through the inhibitory feedback to obtain an inhibitory feedback neural network; and recombining the inhibitory feedback neural network into a closed-loop neural network, and determining a decision rule for feedback learning of the current sample. According to the method, the performance of the whole neural network can be optimized by introducing the concepts of excitability and inhibitory neurons in the neuroscience; and furthermore,the robustness can be improved by introducing a closed-loop system after inhibition feedback.

Description

[0001] This application claims priority to a Chinese patent application with application number 201811341201.2 and titled "Artificial Neural Network Inspired by Brain Inhibitory Feedback Mechanism" filed with the China Patent Office on November 12, 2018, the entire contents of which are hereby incorporated by reference In this application. technical field [0002] The invention belongs to the technical field of artificial neural network and brain neuroscience in computer science, and in particular relates to an artificial neural network construction method and system inspired by the inhibition feedback mechanism of the brain. Background technique [0003] Artificial neural network is a technology that emerged in the field of artificial intelligence in the middle and late last century. To a certain extent, it abstracts the human brain neuron cells mathematically, establishes a neuron model in the sense of information science, and forms different network structures according t...

AI Technical Summary

Problems solved by technology

Later, Minsky, one of the main founders of artificial intelligence, carefully analyzed the functions and limitations of the neural network system represented by the perceptron, and published the book "Perceptron" in 1969, pointing out that the perceptron cannot realize the XOR logic function problem. , and cannot solve the higher-order predicate problem

These studies directly negated the early artificial neural networks, and had a great negative impact on the development of artificial neural networks in the following years.

Images