深度学习驱动的不良信息语义挖掘及快速识别研究

Research on semantic mining and fast recognition of bad information driven by deep learning

网络中的不良信息对用户体验造成重大损害,也对社会政治稳定带来负面影响。电信运营商作为网络服务提供者,应履行政治、社会和商业责任,积极采取措施,打击治理不良信息。本文采用深度学习技术处理大规模数据,通过学习信息文本和通信模式特征,快速自动识别和分类不良信息。此成果还可快速扩展至图片和语音等多媒体不良信息识别,助力开发更全面、智能的不良信息识别系统,为不良信息治理奠定坚实基础。

Bad information in the network causes significant damage to user experience and may also have a negative impact on social and political stability.As network service providers,operators should fulfill their political,social and commercial responsibilities and actively take measures to combat undesirable information.This study uses deep learning technology to process large-scale data and quickly and automatically identify and classify bad information by learning information text and communication mode characteristics to improve network security and user experience.The results can also be rapidly expanded to the recognition of multimedia bad information such as pictures and voice,helping to develop a more comprehensive and intelligent bad information recognition system,and laying a solid foundation for bad information governance.