Neural network construction method and device
A neural network and construction method technology, applied in the field of neural network construction methods and devices, to achieve the effect of improving generalization and increasing probability
Pending Publication Date: 2020-11-13
HUAWEI TECH CO LTD
View PDF0 Cites 6 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Therefore, when building a neural network, after the structural parameters and weight parameters are updated, the discretization basic operation is performed, and the final output network is only obtained by stacking the same building units, which can only be applied to simple scenarios
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0065] The following will describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the scope of protection of this application.
[0066]The neural network construction method provided in this application can be applied in artificial intelligence (artificial intelligence, AI) scenarios. AI is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is the branch of computer science that att...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a neural network construction method and device in the field of artificial intelligence, the method and the device are used for accurately and efficiently constructing a targetneural network, are high in output accuracy of a constructed output network, can be applied to different application scenes, and are high in generalization capability. The method comprises the stepsthat an initial super network is acquired, the initial super network comprises multiple nodes, and the multiple nodes are connected through at least one basic operation; at least one iterative updateis performed on the super network through a preset training set, at least one output network is obtained, wherein in any one iterative update of the at least one iterative update, the at least one iterative update is updated; and the weight parameter and the structure parameter of the output network obtained by the last iterative update are updated through a preset training set to obtain an updated network, and the basic operation corresponding to the structure parameter lower than a preset value in the updated network is deleted to obtain the output network of the current iteration.
Description
technical field [0001] The present application relates to the field of artificial intelligence, in particular to a neural network construction method and device. Background technique [0002] In the field of artificial intelligence, neural networks, especially deep neural networks, have achieved great success in the field of computational vision in recent years. Benefiting from the increase in computing power and the proposal of more and more different components, the neural network structure is developing in a more and more complex direction. [0003] When constructing a neural network, a given hypernetwork is first artificially designed, which may include multiple nodes, and each node is connected by one or more basic operations. Then the hypernetwork is optimized by the bi-level optimization method, and the weight parameters and structural parameter updates are optimized on the training data set and the verification data set respectively until the end of the search, and ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/04G06N3/08
CPCG06N3/082G06N3/045
Inventor 毕恺峰魏龙辉陈鑫谢凌曦田奇
Owner HUAWEI TECH CO LTD