Mongolian rumor detection method based on subjective and objective interpretable bidirectional graph neural network

Abstract

A subjective and objective interpretable Mongolian rumor detection method of a bidirectional graph neural network uses an evidence retrieval module to retrieve evidence, including document retrieval, evidence retrieval and statement verification; constructing two heterogeneous graphs, namely a bidirectional tree graph and an evidence star graph, wherein the bidirectional tree graph learns a rumor propagation mode by using a graph convolutional neural network with a rumor propagation directed graph from top to bottom; capturing a rumor diffusion mode by using a graph convolutional neural network with a rumor diffusion graph from bottom to top; in the evidence star graph, a source post, namely a detection object, is positioned in the center, and evidence nodes surround the source post, so that an angle in the star structure is represented; constructing a rumor detection model based on the bidirectional tree diagram; after K iterations of information transmission based on a bidirectional tree structure and a star-shaped structure, obtaining final representations of two embedding results, connecting the final representations, and transmitting the final representations to a multilayer perceptron for final prediction; mongolian rumor detection can be carried out by utilizing the model, and classification labels and evidence interpretation can be obtained.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, in particular to a Mongolian rumor detection method with a subjective and objectively interpretable bidirectional graph neural network. Background technique [0002] Rumor detection is known as an indispensable part of purifying the Internet ecology. Existing rumor detection methods include artificial rumor detection methods, machine learning-based rumor detection methods, and deep learning-based rumor detection methods. [0003] In recent years, with the rapid development of deep learning, various model structures emerge one after another. Compared with the widely mentioned tweet depth propagation pattern in rumor detection, the widely dispersed structure of rumors is often ignored; in addition, existing rumor detection strategies usually provide detection labels, ignore their explanations, but provide evidence to It is essential to explain why the questionable article is a rumo...

AI Technical Summary

Problems solved by technology

Therefore, the propagation mode analysis and interpretability in its rumor detection will face more difficulties.

Images