Deep learning autonomous operation method

Abstract

The invention discloses a deep learning autonomous operation method. The method comprises the following steps: obtaining an original corpus text; basic semantic segmentation; named entity identification and entity relationship extraction; establishing a basic knowledge graph; establishing an extended knowledge graph; designing a training model and generating training data and a test data set; generating a prediction model. According to the method, the long text information can be automatically stored. Meanwhile, due to the fact that the long text learning and the application program are of anasynchronous structure, the learning process does not affect operation of the application program, independent industry knowledge bases can be formed for different industries, and the related industries comprise the financial industry, the insurance industry, the education industry, the medical industry, the legal industry, the tourism industry, the security industry and the customer service industry.

Description

technical field [0001] The present invention relates to the technical field of deep learning, in particular to a method for autonomous operation of deep learning. Background technique [0002] In 2006, Geoffrey Hinton, a professor at the University of Toronto in Canada and a leader in the field of machine learning, and his student Ruslan Salakhutdinov published an article in the top academic journal "Science", which opened the wave of deep learning in academia and industry. This article has two main messages: 1. Many artificial neural networks with hidden layers have excellent feature learning capabilities, and the learned features have a more essential description of the data, which is conducive to visualization or classification; 2. Deep neural networks The difficulty in training can be effectively overcome by "Layer-wise Pre-training". In this article, layer-wise initialization is achieved through unsupervised learning. [0003] Since 2006, deep learning has continued to...

AI Technical Summary

Problems solved by technology

For example: fruit classification, the feature can choose shape (where the color difference is large, the outline can be determined), color, etc., but it is difficult to find features to distinguish more complex pictures, people looking for features may lack features, or find invalid features, deep learning: That is, the deep neural network. Advantages: no need to manually extract features. Disadvantages: It is impossible to deduce which factors (features) have affected the prediction results, and the features are all shielded in the neural network and cannot be seen by the outside world.

Images