METHOD AND APPARATUS FOR ANALYZING DATA USING ARTIFICIAL NEURAL NETWORK MODEL

摘要

A data analysis method using an artificial neural network model is provided. A data analysis method using an artificial neural network model according to the present invention includes obtaining data to be analyzed, and outputting characteristics of the data to be analyzed by analyzing the data to be analyzed through an artificial neural network model, The neural network model includes a first layer, a first sub-neural network that receives output data of the first layer as input data, and repeatedly includes a first function layer and a pooling layer, and the first and a second sub-neural network that receives the output data of the layer as input data and operates independently of the first sub-neural network without being affected by the operation result of the first sub-neural network. According to this method, object information of various types and sizes can be effectively encoded through several independent sub-neural networks with different filter sizes, and computational efficiency is improved by reducing the number of model parameters and the amount of floating-point computation required for computation. It can also alleviate the problem of overfitting. In addition, since it is easy to process the operation of each independent sub-neural network through mutually independent individual computing devices such as individual cores in a multi-core processor, parallel processing of the artificial neural network model can be easily implemented.