基于区块链和深度学习的全媒体虚假信息检测方法研究

A study of the method of all-media fake information detection based on blockchain and deep learning

在当前的全媒体数字信息环境下,虚假文本,包括假新闻、篡改内容和各类虚假信息,越来越受到关注。为了应对这一挑战,本文引入了一种创新方法,将区块链技术的强大功能与深度学习技术相结合,用于检测虚假文本。该方法构建了一个可信的信息传播网络,保证了信息的安全和透明共享;利用集成的词向量表示来有效地保存内容之间的细粒度关系;采用GRU-CNN模型对输入数据进行训练,训练结果反馈给集成媒体区块链网络。最后的实验结果表明,本文所提出方法在全媒体环境下以短视频为代表的内容任务中检测虚假评论是可行和有效的。

Within the environment of all-media digital information at present,fake texts,including fake news,falsified content and all types of fake information,are more and more widely focused.Replying to this challenge,in this paper a new method for the detection of fake text was introduced by the combination of the power of the blockchain technology with deep learning.With the proposed method,a trustworthy information communication network that ensured the security and transparent sharing of information was built;by using the integrated presentation of word vectors,the fine-grained relations within the contents was stored;via the training of input data with GRU-CNN,the results were fed to the integrated media blockchain network.The final experiment results indicate that the proposed method is feasible and effective in the detection of fake comments in the typical content tasks of short video in the environment of all-media.