杜刚, 朱艳云, 张晨, et al. Research on translation technology of variant spam messages[J]. 2020, 33(7): 83-88. DOI: 10.13992/j.cnki.tetas.2020.07.017.
变体垃圾短信翻译技术研究
摘要
变体垃圾短信被赌博类垃圾信息广泛使用
该类短信使用同音字或形近字替换
绕过垃圾短信监控系统的关键字审查。本文对变体垃圾短信的特点进行了深入研究
并结合人工智能技术
提出了有效翻译变体垃圾短信的技术方法
并给出了具体的应用方案。实验证明
本文提出的变体垃圾短信翻译方法能够对敏感关键词进行完整恢复
便于监控系统对内容进行关键字审查。
Abstract
Variant spam messages are widely used by gambling spam messages. It uses homophone replacement and near-form replacement to bypass keyword censorship in the spam monitoring system. In this paper
the characteristics of variant spam messages are studied in depth
combined with artifi cial intelligence technology
an effective translation technique of variant spam messages is proposed
and a specific application scheme is given. The experiment proves that the variant spam message translation method proposed in this paper can completely recover many sensitive keywords
which is convenient for the monitoring system to conduct keyword review of the content.