Automatic Conformal Representation of Multimodal Evolutionary Features Based on Dynamic Hypergraph Networks

Abstract

The invention belongs to the field of automatic learning of big data machines, and specifically relates to an automatic conformal representation method, system and device for multimodal evolution features based on a dynamic hypergraph network, aiming to solve the problem that the existing multimodal feature representation can only be used for multimodal Static feature representation of dynamic data is difficult to characterize multimodal high-order dynamic correlation features. The method includes obtaining m modal data stream training samples as an input data stream; extracting a finite node set of m modal eigenvectors in the input data stream; generating a Laplacian matrix of m modal hypergraphs; Solve and calculate the high-order correlation conformal entropy between m modes of hypergraphs, and generate a multi-modal high-order dynamic correlation morphological alignment model of n topic network sets; for newly added hypergraph vertices and multi-modal high-order dynamic correlations The nodes covered by the morphological alignment model are added and subtracted to achieve automatic update of the subject network collection. The invention solves the problem of difficult characterization of multi-modal high-order big data multi-element correlation evolution features.

Description

technical field [0001] The invention belongs to the field of automatic learning of big data machines, and in particular relates to an automatic conformal representation method, system and equipment for multi-modal evolution features based on a dynamic hypergraph network. Background technique [0002] With the integration of information technologies such as the Internet, cloud computing, and artificial intelligence, and industries such as manufacturing, transportation, urban management, and medical care, the interaction between people and people, and between people and things is facilitated, and text, image, video, etc. The rapid expansion of information such as audio and audio, along with the Internet of Things such as radar, infrared, multimedia sensors, and infrastructure perception systems, has been continuously integrated into social computing systems, triggering rapid data growth, forming a multi-source heterogeneous, fast, Massive data collections that are difficult to...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above-mentioned problems in the prior art, that is, in order to solve the problem that it is difficult to express multi-modal high-order dynamic associations in the existing multi-modal static feature representation, the first aspect of the present invention proposes a dynamic hypergraph based An automatic conformal representation method for multi-modal evolution features of the network, the method includes:

Images