Machine learning method and machine learning device

Abstract

The invention provides a machine learning method and a machine learning device. The machine learning method includes the steps: constructing an original input matrix and an original output matrix according to an acquired training sample set; constructing a mapping feature node matrix based on the original input matrix by using a first random weight and a first random number; constructing an enhanced node matrix based on the mapping feature node matrix by using a second random weight and a second random number; and according to an augmented matrix composed of the mapping feature node matrix andthe enhanced node matrix and the original output matrix, determining to connect the weight matrix. In the machine learning method and the machine learning device of the present invention, not only the factor of direct connection between the input layer and the output layer in the neural network is added, but also more factors of the hidden layer between the input layer and the output layer is added, so that the machine learning method and the machine learning device have the advantages of satisfying the computational efficiency and guaranteeing the learning accuracy of a wide learning method.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to a machine learning method and a machine learning device. Background technique [0002] Deep structured neural networks and learning have been applied in many fields and have achieved breakthrough success in many applications, especially in large-scale data processing. Among them, the most popular deep networks are Deep Belief Networks (DBN), Deep Boltzmann Machines (DBM) and Convolutional Neural Networks (CNN). Even though deep structured networks are so powerful, most of them suffer from an extremely time-consuming training process due to the complexity of the above deep structures and the large number of hyperparameters involved. Furthermore, this complexity makes it difficult to theoretically analyze the deep structure. In order to obtain higher accuracy in the application, the model has to increase the number of network layers or adjust the number of parameters. Rece...

AI Technical Summary

Problems solved by technology

However, none of these methods can achieve the optimal efficiency and accuracy.

Images