The use of hidden conditional random fields （HCRFs） for tone modeling is explored. The tone recognition performance is improved using HCRFs by taking advantage of intra-syllable dynamic, inter-syllable dynamic and d...The use of hidden conditional random fields （HCRFs） for tone modeling is explored. The tone recognition performance is improved using HCRFs by taking advantage of intra-syllable dynamic, inter-syllable dynamic and duration features. When the tone model is integrated into continuous speech recognition, the discriminative model weight training （DMWT） is proposed. Acoustic and tone scores are scaled by model weights discriminatively trained by the minimum phone error （MPE） criterion. Two schemes of weight training are evaluated and a smoothing technique is used to make training robust to overtraining problem. Experiments show that the accuracies of tone recognition and large vocabulary continuous speech recognition （LVCSR） can be improved by the HCRFs based tone model. Compared with the global weight scheme, continuous speech recognition can be improved by the discriminative trained weight combinations.展开更多
A new speech recognition method is proposed, that integrates a VQ distortion measure and a discrete HMM. This VQ distortion based HMM uses a VQ distortion measure at each state instead of a discrete probability out...A new speech recognition method is proposed, that integrates a VQ distortion measure and a discrete HMM. This VQ distortion based HMM uses a VQ distortion measure at each state instead of a discrete probability output used by a discrete HMM. Although this method is regarded as a refined version of the VQ distortion based recognition method proposed by Burton et al, it is also considered as a special case of a mixed distribution density HMM. In this paper, the VQ distortion based HMM is described, and it is compared with the conventional HMMs and their speech recognition performance through the experiments on speaker independent spoken digit recognition. From these comparisons, we confirm that the new method is better than the traditional HMMs.展开更多
Two discriminative methods for solving tone problems in Mandarin speech recognition are presented. First, discriminative training on the HMM （hidden Markov model） based tone models is proposed. Then an integration t...Two discriminative methods for solving tone problems in Mandarin speech recognition are presented. First, discriminative training on the HMM （hidden Markov model） based tone models is proposed. Then an integration technique of tone models into a large vocabulary continuous speech recognition system is presented. Discriminative model weight training based on minimum phone error criteria is adopted aiming at optimal integration of the tone models. The extended Baum Welch algorithm is applied to find the model-dependent weights to scale the acoustic scores and tone scores. Experimental results show that tone recognition rates and continuous speech recognition accuracy can be improved by the discriminatively trained tone model. Performance of a large vocabulary continuous Mandarin speech recognition system can be further enhanced by the discriminatively trained weight combinations due to a better interpolation of the given models.展开更多
Security of the quantum secure direct communication protocol （i.e., the C-S QSDC protocol） recently proposed by Cao and Song [Chin. Phys. Lett. 23 （2006） 290] is analyzed in the case of considerable quantum channe...Security of the quantum secure direct communication protocol （i.e., the C-S QSDC protocol） recently proposed by Cao and Song [Chin. Phys. Lett. 23 （2006） 290] is analyzed in the case of considerable quantum channel noise. The eavesdropping scheme is presented, which reveals that the C-S QSDC protocol is not secure if the quantum bit error rate （QBER） caused by quantum channel noise is higher than 4.17%. Our eavesdropping scheme induces about 4.17% QBER for those check qubits. However, such QBER can be hidden in the counterpart induced by the noisy quantum channel if the eavesdropper Eve replaces the original noisy channel by an ideal one. Furthermore, if the QBER induced by quantum channel noise is lower than 4.17%, then in the eavesdropping scheme Eve still can eavesdrop part of the secret messages by safely attacking a fraction of the transmitted qubits. Finally, an improvement on the C-S QSDC protocol is put forward.展开更多
It is widely accepted nowadays that intelligibility is the essential goal for most learners of English,and it is not necessary for them to mimic all aspects of native-speaker English in order to achieve a high level o...It is widely accepted nowadays that intelligibility is the essential goal for most learners of English,and it is not necessary for them to mimic all aspects of native-speaker English in order to achieve a high level of intelligibility.However,the features that are needed in order to make oneself easily understood by listeners from elsewhere remain controversial.The current research focuses on thirteen five-minute recordings of conversations between young speakers of English in central China and an interviewer from Britain,in order to determine which features of their speech gave rise to misunderstandings.It was found that,in the 18 tokens of misunderstanding identified,4 resulted from lexical semantics(22%),3 from Chinese place names(17%),3 from grammar(17%),and 11 from pronunciation(61%)(with some tokens cross-classified).The most common phonological factors giving rise to loss of intelligibility were omission of syllables and simplification of word-initial consonant clusters.展开更多
This work is concerned with identification of systems that are subject to not only measurement noises, but also structural uncertainties such as unmodeled dynamics, sensor nonlinear mismatch, and observation bins. Ide...This work is concerned with identification of systems that are subject to not only measurement noises, but also structural uncertainties such as unmodeled dynamics, sensor nonlinear mismatch, and observation bins. Identification errors are analyzed for their dependence on these structural uncertainties. Asymptotic distributions of scaled sequences of estimation errors are derived.展开更多
文摘The use of hidden conditional random fields （HCRFs） for tone modeling is explored. The tone recognition performance is improved using HCRFs by taking advantage of intra-syllable dynamic, inter-syllable dynamic and duration features. When the tone model is integrated into continuous speech recognition, the discriminative model weight training （DMWT） is proposed. Acoustic and tone scores are scaled by model weights discriminatively trained by the minimum phone error （MPE） criterion. Two schemes of weight training are evaluated and a smoothing technique is used to make training robust to overtraining problem. Experiments show that the accuracies of tone recognition and large vocabulary continuous speech recognition （LVCSR） can be improved by the HCRFs based tone model. Compared with the global weight scheme, continuous speech recognition can be improved by the discriminative trained weight combinations.
文摘A new speech recognition method is proposed, that integrates a VQ distortion measure and a discrete HMM. This VQ distortion based HMM uses a VQ distortion measure at each state instead of a discrete probability output used by a discrete HMM. Although this method is regarded as a refined version of the VQ distortion based recognition method proposed by Burton et al, it is also considered as a special case of a mixed distribution density HMM. In this paper, the VQ distortion based HMM is described, and it is compared with the conventional HMMs and their speech recognition performance through the experiments on speaker independent spoken digit recognition. From these comparisons, we confirm that the new method is better than the traditional HMMs.
文摘Two discriminative methods for solving tone problems in Mandarin speech recognition are presented. First, discriminative training on the HMM （hidden Markov model） based tone models is proposed. Then an integration technique of tone models into a large vocabulary continuous speech recognition system is presented. Discriminative model weight training based on minimum phone error criteria is adopted aiming at optimal integration of the tone models. The extended Baum Welch algorithm is applied to find the model-dependent weights to scale the acoustic scores and tone scores. Experimental results show that tone recognition rates and continuous speech recognition accuracy can be improved by the discriminatively trained tone model. Performance of a large vocabulary continuous Mandarin speech recognition system can be further enhanced by the discriminatively trained weight combinations due to a better interpolation of the given models.
基金The project supported by the Program for New Century Excellent Talents at the University of China under Grant No.NCET-06-0554the National Natural Science Foundation of China under Grant No.60677001+3 种基金the Science Technology Fund of Anhui Province for Outstanding Youth under Grant No.06042087the Key Fund of the Ministry of Education of China under Grant No.206063the Natural Science Foundation of Guangdong Province under Grant Nos.06300345 and 7007806Natural Science Foundation of Hubei Province under Grant No.2006ABA354
文摘Security of the quantum secure direct communication protocol （i.e., the C-S QSDC protocol） recently proposed by Cao and Song [Chin. Phys. Lett. 23 （2006） 290] is analyzed in the case of considerable quantum channel noise. The eavesdropping scheme is presented, which reveals that the C-S QSDC protocol is not secure if the quantum bit error rate （QBER） caused by quantum channel noise is higher than 4.17%. Our eavesdropping scheme induces about 4.17% QBER for those check qubits. However, such QBER can be hidden in the counterpart induced by the noisy quantum channel if the eavesdropper Eve replaces the original noisy channel by an ideal one. Furthermore, if the QBER induced by quantum channel noise is lower than 4.17%, then in the eavesdropping scheme Eve still can eavesdrop part of the secret messages by safely attacking a fraction of the transmitted qubits. Finally, an improvement on the C-S QSDC protocol is put forward.
基金funded by State Administration of Foreign Expert Affairs(project GDT20173200030&project G20190214022).
文摘It is widely accepted nowadays that intelligibility is the essential goal for most learners of English,and it is not necessary for them to mimic all aspects of native-speaker English in order to achieve a high level of intelligibility.However,the features that are needed in order to make oneself easily understood by listeners from elsewhere remain controversial.The current research focuses on thirteen five-minute recordings of conversations between young speakers of English in central China and an interviewer from Britain,in order to determine which features of their speech gave rise to misunderstandings.It was found that,in the 18 tokens of misunderstanding identified,4 resulted from lexical semantics(22%),3 from Chinese place names(17%),3 from grammar(17%),and 11 from pronunciation(61%)(with some tokens cross-classified).The most common phonological factors giving rise to loss of intelligibility were omission of syllables and simplification of word-initial consonant clusters.
文摘This work is concerned with identification of systems that are subject to not only measurement noises, but also structural uncertainties such as unmodeled dynamics, sensor nonlinear mismatch, and observation bins. Identification errors are analyzed for their dependence on these structural uncertainties. Asymptotic distributions of scaled sequences of estimation errors are derived.
