摘要
探讨了运营商智能客服的技术发展趋势,介绍了自动语音识别(automatic speech recognition,ASR)转写纠错技术、语义提取技术、加密数据库以及坐席访问控制等技术。分析了语音识别的错误检测和自动纠错两种不同的研究方向;在语义理解方面分析了小样本训练下的有监督学习和无监督的关键词提取;介绍了大数据加密和坐席访问控制技术。同时,展望了多模态交互技术、智能推荐技术和面向残疾人服务等技术方向。总之,智能客服技术的发展与创新将为通信行业带来更高效、便捷的服务体验,推动行业服务水平的进步。
The technology development trend of intelligent customer service for operators was discussed,and automatic speech recognition(ASR)transcription error correction technology,semantic extraction technology,encrypted database,and agent access control technology were introduced.Two different research directions of error detection and automatic error correction for speech recognition were analyzed,two directions of supervised learning and unsupervised learning were discussed,and big data encryption and agent access control technologies were introduced.At the same time,the technical directions of multimodal interaction technology,intelligent recommendation technology and disabled-oriented services were prospected.In conclusion,the development and innovation of intelligent customer service technology will bring a more efficient and convenient service experience to the communication industry and strongly promote the progress of the industry service level.
作者
马晓亮
刘英
杜德泉
安玲玲
MA Xiaoliang;LIU Ying;DU Dequan;AN Lingling(Xidian University,Xi’an 710126,China;Guangzhou Branch of China Telecom Co.,Ltd.,Guangzhou 510620,China;Ma Xiaoliang Innovation Studio for Model Workers and Creative Talents,Guangzhou 510620,China)
出处
《电信科学》
2023年第5期76-89,共14页
Telecommunications Science
基金
国家重点研发计划项目(No.2022YFB3102700,No.2018YFB1802400)
国家自然科学基金重点项目(No.62132013)。
