Convolutional neural network compression method and system based on learning automaton and medium

Abstract

The invention provides a convolutional neural network compression method and system based on a learning automaton and a medium. The method comprises a parameter initialization step of initializing theparameters of the learning automaton; a state value selection step of according to the obtained initialized learning automaton parameters, enabling each learning automaton to select a state value ofthe learning automaton according to a preset behavior selection probability to obtain a state value of each learning automaton; and a network structure updating step of updating the network structureaccording to the obtained state value of each learning automaton, and obtaining the updated network structure. According to the method, the learning automaton idea is innovatively used for screening the optimal convolution kernel set in the convolutional neural network, so that the convolutional neural network can complete the network compression task to the maximum extent under the condition of losing a little of classification precision.

Description

technical field [0001] The present invention relates to the fields of deep learning and artificial intelligence, in particular to a convolutional neural network compression method, system and medium based on learning automata. In particular, it relates to a convolutional neural network compression method based on Learning Automata (LA). Background technique [0002] Convolutional neural network is a mathematical model that can simulate complex functions. It belongs to the feedforward neural network in deep learning. This model gradually converts the initial features of the input data into advanced features through a multi-layer network, and then combined with subsequent related operations can Complete complex classification learning tasks. [0003] Convolutional neural network is currently mainly used in the field of image processing. It imitates the human visual processing process through multi-layer convolution and pooling. The effect of scale transformation is to form m...

AI Technical Summary

Problems solved by technology

Both have different problems in practical applications, among which the parameter compression effect of encoding compression is small, and the pruning based on weight does not consider the correlation between connections, and the loss of accuracy is relatively large to a certain extent.

Therefore, how to efficiently compress convolutional neural networks is still a huge challenge

Images