Social text dependency syntactic analysis system based on deep neural network

Abstract

A social text dependency syntax analysis system based on a deep neural network relates to the technical field of computer information processing and aims to the problem of sparse social text data in the prior art. The system comprises a social text crawling module, a preprocessing module, a base bilinear attention module, a stack bilinear attention module and a joint decoding and training module.The social text crawling module is used for crawling a social text from a social media website; the preprocessing module is used for filtering the obtained social text and generating an initializationword vector; the base bilinear attention module is used for performing pre-training by utilizing a regular text; the stack type bilinear attention module is used for predicting a social text; and thejoint decoding and training module is used for calculating an experience risk function for the stack type bilinear attention module, performing back propagation gradient parameter adjustment, fittinga training function, and finally performing parallel computing acceleration model training by utilizing a GPU.

Description

technical field [0001] The invention relates to the technical field of computer information processing, in particular to a deep neural network-based social text dependency syntax analysis system. Background technique [0002] Dependency analysis is a basic and important task in natural language processing. Many applications require dependency analysis on sentences to provide syntactic results for corresponding tasks. Through the powerful computing power of the computer, the dependent syntactic structure of the sentence is identified. The dependency syntax tree is roughly divided into two types according to the structure: projective (Project) and non-projective (Non-project) dependency syntactic structure; according to the decoding algorithm: Graph-based and transition-based (Transition-based) dependencies algorithm. The deep neural network partially overcomes the gradient dispersion and explosion of the traditional neural network, and has developed rapidly in recent years,...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to propose a social text dependency syntax analysis system based on a deep neural network for the problem of sparse social text data in the prior art

Images