中国计算语言学研究现状与展望  

Research Status and Prospects of Computational Linguistics

作　　者：耿立波 酆格斐 詹卫东[4] 杨亦鸣 Geng Libo;Feng Gefei;Zhan Weidong;Yang Yiming(School of Linguistic Sciences and Arts,Jiangsu Normal University,Xuzhou Jiangsu 221009;Jiangsu/China Collaborative Innovation Center for Language Ability;Xuzhou Jiangsu 2210093Jiangsu Key Laboratory of Language and Cognitive Neuroscience,Xuzhou Jiangsu 221009;MOE Key Laboratory of Computational Linguistics,Peking University,Beijing 100871)

机构地区：[1]江苏师范大学语言科学与艺术学院,江苏徐州221009 [2]语言能力省部共建协同创新中心,江苏徐州221009 [3]江苏省语言与认知神经科学重点实验室,江苏徐州221009 [4]北京大学计算语言学教育部重点实验室,北京100871

出　　处：《语言科学》2021年第5期491-499,共9页Linguistic Sciences

基　　金：国家社科基金重大委托项目研究专项(19VXK06)和国家社科基金青年项目(16CYY021)资助。

摘　　要：“十三五”期间我国的计算语言学取得了长足的进步与发展,受到深度学习算法的推进,基础研究方面出现了较大突破,在语音识别、机器翻译、自动问答系统、知识资源建设、古文字和其他语种语言信息处理等应用方面也出现了很多重要成果。但与世界先进水平相比,目前在很多领域内我们还只是处于跟跑阶段,并且深度学习算法的红利也已接近释放殆尽,在未来仍需要从算法基础架构、人脑语言的本质、深层语言理解等方面展开研究,发展机器语言能力等新兴方向,并积极开展复合型语言学人才的培养。Computational linguistics had made considerable progress in China during the period of the“13th Five-Year Plan”.Inspired by deep learning algorithms,certain breakthroughs have been made in the basic research aspects.In the meantime,this period also witnessed valuable applications in computational linguistics in areas such as speech recognition,machine translation,automatic question and answer systems,knowledge resource construction,as well as language and information processing of ancient Chinese characters and other languages.However,we are still catching up in many fields.Moreover,the dividends brought by deep learning algorithms are draining away.To win the competition,we need to conduct basic research on algorithms,the neural mechanism of language,and the nature of language comprehension.We also need to invest in training interdisciplinary talents and strengthen research in emerging areas such as machine language ability.

关 键 词：计算语言学 深度学习 机器语言能力 

分 类 号：H087[语言文字—语言学]