| 432 | 11 | 439 |
| 下载次数 | 被引频次 | 阅读次数 |
近年来,随着深度学习和语音识别技术的飞速发展,基于深度学习语音识别的计算机辅助外语口语学习成为当前人工智能技术应用研究的一个热点。本文结合当前最先进的智能语音信息处理理论,在阐述英语口语自动评测的基本原理和算法的基础上,针对中考、高考口语考试考生音频的特点,提出了两种基于深度神经网络声学模型的更具噪音鲁棒性的评分算法。依据在初中和高中英语口语大规模统一考试的真实场景数据进行的验证实验,本文提出的自动评测方法比传统基于GOP(Goodness of Pronunciation)的方法具有较大的性能优势。本研究开发的部分技术已实际应用于全国多地的中考、高中期末考试及高考模拟考试的口语自动阅卷系统中,取得了良好的社会效益。
Abstract：In recent years, with the rapid development of deep learning and speech recognition technology, computer-assisted spoken language learning based on in-depth learning speech recognition has become a hot topic in the application of artificial intelligence technology. Combined with the most advanced theory of intelligent speech information processing, this paper expounds the basic principles and algorithms of the automatic evaluation of spoken English, and aims at the characteristics of the audio frequency of the examinees in the oral examinations of the middle school entrance examination and the college entrance examination. Two more robust scoring algorithms based on depth neural network acoustic model are proposed. Based on the real scene data from the large scale unified test of spoken English in junior high school and senior high school, the automatic evaluation method proposed in this paper has greater performance advantages than the traditional GOP(good of Pronunciation) method. Some of the techniques developed in this study have been applied to the oral automatic marking system of middle school entrance examination, senior middle school final examination and college entrance examination simulation examination in many places of China, and have achieved good social benefits.
[1]Cheng,J.Automatic Assessment of prosody in high-stakes English tests[J],INTERSPEECH 2011,1589-1592.
[2]Dean Luo,et al.Investigation of the Effects of Automatic Scoring Technology on Human Raters’Performances in L2Speech Proficiency Assessment[J],ISCSLP,2016,1-4.
[3]D.Luo,Y.Qiao,N.Minematsu,Y.Yamauchi,K.Hirose,Analysis and utilization of MLLR speaker adaptation technique for learners pronunciation evaluation,Proc.INTERSPEECH,2009(9)608-611.
[4]D.Luo,Y.Qiao,N.Minematsu,Y.Yamauchi,K.Hirose,Regularized-MLLR Speaker Adaptation for Computer-Assisted Language Learning System[J],Proc.INTERSPEECH.2010(9)594-597.
[5]Dean Luo,Naoya Shimomura,Nobuaki Minematsu,Yutaka Yamauchi and Keikichi Hirose,Automatic Pronunciation Evaluation of Language Learners’Utterances Generated through Shadowing[J],Proc.INTERPEECH.2008(9)2807-2810.
[6]C.Tsurutani,Y.Yamauchi,N.Minematsu,D.Luo,K.Maruyama,and K.Hirose,Development of a program for self assessment of Japanese pronunciation by English learners,Proc.ICSLP’2006(9),841-844.
[7]D.Luo,R.Luo,and L.Wang,Naturalness Judgement of L2English Through Dubbing Practice[J],Proc.Interspeech2016,200-203
[8]http://www.seaskylight.com/gsxw/info_24.aspx?itemid=2102
[9]S.M.Witt and S.J.Young.(2000).Phone-level Pronunciation Scoring and Assessment for Interactive Language Learning,”Speech Communications,30(2-3):pp.95-108.
[10]W.Hu,et al.(2013),A New DNN-based High Quality Pronunciation Evaluation for Computer-Aided Language Learning(CALL),Proc.INTERSPEECH 2013,1886-1890
[11]D.Povey,A.Ghoshal,G.Boulianne,L.Burget,O.Glembek,N.Goel,M.Hannemann,P.Motlicek,Y.Qian,P.Schwarz,J.Silovsky,G.Stemmer,and K.Vesely,“The Kaldi speech recognition toolkit,”in Proc.ASRU,2011.
[12]Panayotov,V.,Chen,G.,Povey,D.,&Khudanpur,S.(2015).Librispeech:An ASR corpus based on public domain audio books.IEEE International Conference on Acoustics,Speech and Signal Processing(pp.5206-5210).IEEE.
[13]Povey,D.,Cheng,G.,Wang,Y.,Li,K.,Xu,H.,Yarmohamadi,M.,Khudanpur,S.:Semi-orthogonal lowrank matrix factorization for deep neural networks.In:INTERSPEECH(2018-submitted)
基本信息:
中图分类号:G633.41
引用信息:
[1]罗德安,夏林中,张春晓,等.基于深度学习的中学生英语口语自动评测技术[J].深圳信息职业技术学院学报,2018,16(02):100-104.
基金信息:
深圳市科技计划项目(GRCK2017042409560810);深圳市科技计划项目(GRCK2017042409552883); 深圳信息职业技术学院科研平台培育项目(PT201701)
2018-06-15
2018-06-15