Filters








806 Hits in 7.6 sec

Automatic Detection of Accent and Lexical Pronunciation Errors in Spontaneous Non-Native English Speech

Konstantinos Kyriakopoulos, Kate M. Knill, Mark J.F. Gales
2020 Interspeech 2020  
Detecting individual pronunciation errors and diagnosing pronunciation error tendencies in a language learner based on their speech are important components of computer-aided language learning (CALL).  ...  Three annotated corpora of non-native English speech by speakers of multiple L1s are analysed, the consistency of human annotation investigated and a method presented for detecting individual accent and  ...  Instead, an ASR system trained on non-native learners of English is used [36, 37] .  ... 
doi:10.21437/interspeech.2020-2881 dblp:conf/interspeech/KyriakopoulosKG20 fatcat:vfsdev64cza5bou2b2uv5mjhr4

Impact of ASR Performance on Free Speaking Language Assessment

Kate Knill, Mark Gales, Konstantinos Kyriakopoulos, Andrey Malinin, Anton Ragni, Yu Wang, Andrew Caines
2018 Interspeech 2018  
This form of test allows the spoken language proficiency of a non-native speaker of English to be assessed more fully than read aloud tests.  ...  Then, the impact of ASR errors on how well the system can detect whether a learner's answer is relevant to the question asked is evaluated.  ...  Both read aloud and free speaking tasks require the ASR to be capable of recognising non-native English speech. Across This research was funded under the ALTA Institute, Cambridge University.  ... 
doi:10.21437/interspeech.2018-1312 dblp:conf/interspeech/KnillGKMRWC18 fatcat:rrdrv6y3ira47dz75kc7ugtite

Pronunciation scoring for Indian English learners using a phone recognition system

Chitralekha Bhat, K. L. Srinivas, Preeti Rao
2010 Proceedings of the First International Conference on Intelligent Interactive Technologies and Multimedia - IITM '10  
Well-known limitations in the accuracy of automatic speech recognition (ASR) systems pose challenges to the reliable detection of pronunciation errors in the speech of non-native speakers.  ...  The system is evaluated on Indian English speech in the realistic situation where there is no matching database available for training the speech recognizer.  ...  To control for the effects of speech recognition errors, it is important to constrain the recognition process by limiting the output phone sequences based on some knowledge of non-native speaking errors  ... 
doi:10.1145/1963564.1963587 fatcat:fr76phqo3fdjnkeuxlw7hoj4m4

Cross-Lingual Transfer Learning of Non-Native Acoustic Modeling for Pronunciation Error Detection and Diagnosis

Richeng Duan, Tatsuya Kawahara, Masatake Dantsuji, Hiroaki Nanjo
2019 IEEE/ACM Transactions on Audio Speech and Language Processing  
In this work, we address non-native acoustic modeling (both on phonetic and articulatory level) based on transfer learning.  ...  It also improves the pronunciation error detection based on goodness of pronunciation (GOP) score.  ...  In the non-native speech recognition and pronunciation error detection experiments, we confirmed the effectiveness of the proposed cross-lingual based transfer learning on acoustic-phonetic modeling.  ... 
doi:10.1109/taslp.2019.2955858 fatcat:3k6x4qtjyvdn7ghkdeix3tuvb4

Open Challenge for Correcting Errors of Speech Recognition Systems [article]

Marek Kubis, Zygmunt Vetulani, Mikołaj Wypych, Tomasz Ziętkiewicz
2020 arXiv   pre-print
The goal of the challenge is to investigate methods of correcting the recognition results on the basis of previously made errors by the speech processing system.  ...  The paper announces the new long-term challenge for improving the performance of automatic speech recognition systems.  ...  Hypotheses -textual outputs of the automatic speech recognition system. 2. References -transcriptions of sentences being read to the automatic speech recognition system.  ... 
arXiv:2001.03041v1 fatcat:ynelpuwqorarhcf6thot5omcxu

Predicting gradation of L2 English mispronunciations using ASR with extended recognition network

Hao Wang, Helen Meng, Xiaojun Qian
2013 2013 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference  
We obtained perceptual gradations of L2 English mispronunciations through crowdsourcing, conducted quality control to filter for reliable ratings and proposed approaches to predict gradation of word-level  ...  This paper presents our work on making improvements using ASR with extended recognition network to the previous predicting approach to solve its limitations: 1. it is not working for those mispronounced  ...  CROWDSOURCED MISPRONUNCIATION GRADATIONS Our previous work [2] collected perceptual gradations of word-level mispronunciations in non-native English speech using the AMT crowdsourcing platform.  ... 
doi:10.1109/apsipa.2013.6694165 dblp:conf/apsipa/WangMQ13 fatcat:t3fr7wflwvgvpl2nmsfkltjrqu

AequeVox: Automated Fairness Testing of Speech Recognition Systems [article]

Sai Sathiesh Rajan
2022 arXiv   pre-print
Our experiments reveal that non-native English, female and Nigerian English speakers generate 109%, 528.5% and 156.9% more errors, on average than native English, male and UK Midlands speakers, respectively  ...  Automatic Speech Recognition (ASR) systems have become ubiquitous. They can be found in a variety of form factors and are increasingly important in our daily lives.  ...  On average, speech from non-native English speakers generates 109% more errors in comparison to speech from native English speakers.  ... 
arXiv:2110.09843v2 fatcat:cfrmcuteejgirixvzita6rlmli

Automatic detection of plagiarized spoken responses

Keelan Evanini, Xinhao Wang
2014 Proceedings of the Ninth Workshop on Innovative Use of NLP for Building Educational Applications  
This paper addresses the task of automatically detecting plagiarized responses in the context of a test of spoken English proficiency for non-native speakers.  ...  ASR output (with a baseline accuracy of 50.0%).  ...  In this paper, we present an initial study of automated plagiarism detection on spoken responses containing spontaneous non-native speech.  ... 
doi:10.3115/v1/w14-1803 dblp:conf/bea/EvaniniW14 fatcat:mts243jr7jeibgie3ydbwnvvri

Error Patterns for Automatic Error Detection in Computer Assisted Pronunciation Training Systems

Olga Kolesnikova
2014 Research in Computing Science  
This paper presents error patterns built on the basis of our comparative analysis of American English and Mexican Spanish phonemes and allophones which can be applied in designing the error detection module  ...  To the best of our knowledge, error patterns in American English speech generated by Mexican Spanish speakers has not been defined in previous work which was done mainly for Castilianoriginated standard  ...  Automatic Speech Recognition in CAPT Systems The basic goal of Automatic Speech Recognition (ASR) is to take an acoustic waveform as input and produce a string of words as output.  ... 
doi:10.13053/rcs-84-1-8 fatcat:qbsykmezergubmjeho4gemgmzy

Extending automatic transcripts in a unified data representation towards a prosodic-based metadata annotation and evaluation

F. Batista, H. Moniz, I. Trancoso, N. Mamede, A. I. Mata
2021 Journal of Speech Sciences  
Initially put to use for automatic detection of punctuation marks and for capitalization recovery from speech data, it has also been recently used for studying the characterization of disfluencies in speech  ...  the previous output in order to accommodate prosodic information.  ...  A Gaussian mixture model (GMM) classifier is then used to automatically detect speech/non-speech regions, based on the energy.  ... 
doi:10.20396/joss.v2i2.15035 fatcat:z7gktxp47rcslkyfs2e3ittxqy

Recognize Mispronunciations to Improve Non-Native Acoustic Modeling Through a Phone Decoder Built from One Edit Distance Finite State Automaton

Wei Chu, Yang Liu, Jianwei Zhou
2020 Interspeech 2020  
This work also offered a data-driven approach for generating a list of common mispronunciation patterns of non-native English learners that may be useful for speech assessment purpose.  ...  The components shown in the diagram are all standard ones except the two rounds of the 'GOP-based GMM alignment' procedure, which will be explained in details in the following sections.  ...  decision tree on native speech. • Step 2: GOP-based Mispronunciation Detection with Native GMM For a non-native spoken utterance and its generated phone sequence according to the word-level transcription  ... 
doi:10.21437/interspeech.2020-3109 dblp:conf/interspeech/ChuLZ20 fatcat:p5l4o6qoebc4vkyorbihyebqnm

Speech Processing for Language Learning: A Practical Approach to Computer-Assisted Pronunciation Teaching

Natalia Bogach, Elena Boitsova, Sergey Chernonog, Anton Lamtev, Maria Lesnichaya, Iurii Lezhenin, Andrey Novopashenny, Roman Svechnikov, Daria Tsikach, Konstantin Vasiliev, Evgeny Pyshkin, John Blake
2021 Electronics  
Both are designed on a base of a third-party automatic speech recognition (ASR) library Kaldi, which was incorporated inside StudyIntonation signal processing software core.  ...  automatically in our code.  ...  In this work we applied voice activity detection (VAD) before pitch processing and instrumented StudyIntonation with a third-party automatic speech recognition (ASR) system.  ... 
doi:10.3390/electronics10030235 fatcat:io6bpmxglbeo5jp4bbz6depy5e

Optimizing Automatic Speech Recognition for Low-Proficient Non-Native Speakers

Joost van Doremalen, Catia Cucchiarini, Helmer Strik
2010 EURASIP Journal on Audio, Speech, and Music Processing  
Computer-Assisted Language Learning (CALL) applications for improving the oral skills of low-proficient learners have to cope with non-native speech that is particularly challenging.  ...  Since unconstrained non-native ASR is still problematic, a possible solution is to elicit constrained responses from the learners.  ...  Improving ASR performance on non-native speech can also be carried out at the level of the lexicon.  ... 
doi:10.1186/1687-4722-2010-973954 fatcat:rjzuc3qrmnfqdnqwgrtchac344

Optimizing Automatic Speech Recognition for Low-Proficient Non-Native Speakers

Joost van Doremalen, Catia Cucchiarini, Helmer Strik
2010 EURASIP Journal on Audio, Speech, and Music Processing  
Computer-Assisted Language Learning (CALL) applications for improving the oral skills of low-proficient learners have to cope with non-native speech that is particularly challenging.  ...  Since unconstrained non-native ASR is still problematic, a possible solution is to elicit constrained responses from the learners.  ...  Improving ASR performance on non-native speech can also be carried out at the level of the lexicon.  ... 
doi:10.1155/2010/973954 fatcat:d2mz4ydn7naffctrhehtadeoda

Evaluating prosodic features for automated scoring of non-native read speech

Klaus Zechner, Xiaoming Xi, Lei Chen
2011 2011 IEEE Workshop on Automatic Speech Recognition & Understanding  
In our first experiment, we compute features based on a positional match between automatically identified stress and tone labels for 741 non-native read text passages with a human gold standard on the  ...  We evaluate two types of prosodic features utilizing automatically generated stress and tone labels for non-native read speech in terms of their applicability for automated speech scoring.  ...  earlier versions of this paper.  ... 
doi:10.1109/asru.2011.6163975 dblp:conf/asru/ZechnerXC11 fatcat:cle7dq6sebbrlfxm5zvtftzdla
« Previous Showing results 1 — 15 out of 806 results