讯飞语音输入法方言识别在新冠疫情防控中的应用评估

汪高武,庞 博,李晨光,杨豆豆

语言战略研究 ›› 2020, Vol. 5 ›› Issue (5) : 48-56.

欢迎访问《语言战略研究》官方网站!今天是
语言战略研究 ›› 2020, Vol. 5 ›› Issue (5) : 48-56. DOI: 10.19689/j.cnki.cn10-1361/h.20200504
语言战略研究

讯飞语音输入法方言识别在新冠疫情防控中的应用评估

  • 汪高武,庞 博,李晨光,杨豆豆
作者信息 +

An Evaluation of Xunfei Speech Input Software in the COVID-19 Pandemic Prevention

  • Wang Gaowu, Pang Bo, Li Chenguang and Yang Doudou
Author information +
文章历史 +

摘要

在各种社会重大突发事件的应急处理中,智能语音技术能起到弥补人力资源严重不足、紧急处理海量信息、有效应对风险、降低损失的作用,具有独特的优势。在2020年新冠肺炎疫情中,全国各地支援湖北,不同方言人群之间的沟通也成为一个问题,这对语音识别技术提出了要求和挑战。本文以科大讯飞智能语音输入法为例,根据全国多种方言(139个方言点,共283位发音人)在疫情防控模拟场景下的语音识别材料,对其成效进行了评估和分析,并探讨了智能语音技术的应用和发展。

Abstract

Intelligent speech technology has a unique advantage in compensating for the serious shortage of human resources, handling huge amounts of information and reducing the risk in a public emergency. In fighting the COVID-19, however, a communication and service problem in domestic emergency assistance has emerged as a challenge for intelligent speech technology, as there are many dialects in China. In this study, Xunfei speech input software, one of the best available in the Chinese market, was evaluated in the context of fighting the COVID-19 Pandemic. A total of 283 participants in 139 dialect communities took part in a simulated medical situation, where their speeches in their local dialects were recognized by the software. The recognition results have been analyzed and discussed, for exploring the prospects for intelligent speech technology to better serve people in a public emergency.

关键词

智能语音技术 / 语音识别 / 汉语方言 / 重大突发事件

Key words

intelligent speech technology / speech recognition / Chinese dialects / public emergency

引用本文

导出引用
汪高武,庞 博,李晨光,杨豆豆. 讯飞语音输入法方言识别在新冠疫情防控中的应用评估[J]. 语言战略研究. 2020, 5(5): 48-56 https://doi.org/10.19689/j.cnki.cn10-1361/h.20200504
Wang Gaowu, Pang Bo, Li Chenguang and Yang Doudou. An Evaluation of Xunfei Speech Input Software in the COVID-19 Pandemic Prevention[J]. Chinese Journal of Language Policy and Planning. 2020, 5(5): 48-56 https://doi.org/10.19689/j.cnki.cn10-1361/h.20200504

基金

本研究由国家自科基金(11974054)和教育部社科青年基金(16YJC740065)资助。感谢匿名评审专家和编辑的宝贵意见!

Accesses

Citation

Detail

段落导航
相关文章

/