DDCAMS:NLP赋能智慧咨询业务  

作　　者：孙甜 耿永建 胡程忆 霍明 杨志强 黄昭 孟龙波 

机构地区：[1]卓望信息技术(北京)有限公司,北京100060

出　　处：《科技创新与应用》2023年第6期68-72,共5页Technology Innovation and Application

摘　　要：新闻网站上的海量新闻具有行业商机、信息洞察等重要研究价值,利用自然语言处理技术进行自动化的信息萃取,替代纯人工筛选信息,方便完成生产报告并推送给领导或关键人。该文以中国移动智慧咨询新闻萃取业务场景为依托,提出DDCAMS,系统介绍从海量新闻当中筛选和处理信息的技术架构及构建流程,包括文本去重、文本去噪、文本分类和文本摘要4个模型,目前已完成初代版本的研发,性能达到预期。打造中国移动AI引领业务变革的应用实践案例,有效提高数智化管理水平,助力公司建设成为“一流的数智化服务提供商”。The massive news on the Internet has important research value such as industry business opportunities and information insight. The use of natural language processing technology for automatic information extraction can replace pure manual screening of information, and it is convenient to complete production reports and push them to leaders or key people. Based on the business scenario of China Mobile Smart Consulting News Extraction, this paper proposes DDCAMS, which systematically introduces the technical architecture and construction process for filtering and processing information from massive news, including text deduplication, text denoising, text classification, and text summarization 4 models, the research and development of the firstgeneration version has been completed, and the performance has reached expectations, so as to create an application practice case of China Mobile’s AI leading business transformation, effectively improve the level of digital intelligence management, and help the company to become a "first-class digital and intelligent service provider".

关 键 词：自然语言处理 文本去重 文本分类 文本摘要 智慧运营 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]