Frame-Level Selective Decoding Using Native and Non-native Acoustic Models for Robust Speech Recognition to Native and Non-native Speech
نویسندگان
چکیده
v Regarded as a mismatch problem between the training and test conditions § Training condition: native speech § Testing condition: non-native speech § Widely used methods in speaker or environment adaptation v Research works dedicated to non-native ASR § Acoustic modeling § Pronunciation modeling § Language modeling § Hybrid modeling § Many researches uses a small amount of non-native speech What to do... v Use a large amount of both native speech and non-native speech v Obtain considerable performances for both native and non-native speech Increased ASR applications Globalization Need of nonnative ASR
منابع مشابه
Speech Recognition of Non-native Speech Using Native and Non-native Acoustic Models
A speech recognition system is subjected to the speech of non-native speakers, using both native and non-native acoustic phone models. The problems involved with the mapping of phoneset from the nonnative to native language are investigated, and a detailed analysis of phone confusions is made. For Dutch speakers, British English acoustic models give the best word recognition results.
متن کاملLikelihood Combination and Recog the Decoding of Non-native Speech
In this paper we report on the combination of multilingual Hidden Markov Models for the recognition of non-native speech. Using a digit recognition task as an example, we first demonstrate the benefits of bilingual acoustic models that incorporate training data from both the target language and the speakers’ native language, and then compare two different recognizer combination methods, namely ...
متن کاملSpeech recognition for multiple non-native accent groups with speaker-group-dependent acoustic models
In this paper, the recognition performance for non-native English speech with two different kinds of speaker-groupdependent acoustic models is investigated. The approaches for creating speaker groups include knowledge-based grouping of non-native speakers by their first language, and the automatic clustering of speakers. Clustering is based on speakerdependent acoustic models in speaker Eigensp...
متن کاملNative and Non-native English Teachers’ Rating Criteria and Variation in the Assessment of L2 Pragmatic Production: The Speech Act of Compliment
Pragmatic assessment and consistency in rating are among the subject matters which are still in need of more profound investigations. The importance of the issue is highlighted when remembering that inconsistency in ratings would surely damage the test fairness issue in assessment and lead to much diversity in ratings. Our principal concern in this study was observing the criteria that American...
متن کاملIntegration of MLLR adaptation with pronunciation proficiency adaptation for non-native speech recognition
To recognize non-native speech, larger acoustic/linguistic distortions must be handled adequately in acoustic modeling, language modeling, lexical modeling, and/or decoding strategy. In this paper, a novel method to enhance MLLR adaptation of acoustic models for non-native speech recognition is proposed. In the case of native speech recognition, MLLR speaker adaptation was successfully introduc...
متن کامل