Enhanced Phone Posteriors for Improving Speech Recognition Systems

Ketabdar, Hamed; Bourlard, Hervé

doi:10.1109/tasl.2009.2023162

Cited by 34 publications

(31 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If this second stage net is trained on several neighboring frames -similar to the MFCC-based net -then it is able to correct some of the errors of the lower stage net(s) with the help of the long-term context. Hence, applying such a second stage network is already useful in itself, as was recently shown in [15] or [16]. We will compare our earlier results with two 2-stage configurations: the first is trained only on the MFCC-based posteriors, while the second combines the MFCC-based and the 2D-DCT based probabilities.…”

Section: Noisy Speech Experimentsmentioning

confidence: 80%

“…Both physiological and psychoacoustic experimental results indicate that the human brain extracts information from much longer time spans. Technically the simplest solution for this is to work with larger windows along the timeaxis: in neural-net based recognizers it is now standard practice to train the system on 9 or more neighboring MFCC vectors [7,15,16]. However, there is also evidence that the brain processes relatively narrow frequency bands quasi-separately [2,6].…”

Section: Localized Spectro-temporal Featuresmentioning

confidence: 99%

See 1 more Smart Citation

Phone recognition experiments with 2D-DCT spectro-temporal features

Kovács

Tóth

2011

2011 6th IEEE International Symposium on Applied Computational Intelligence and Informatics (SACI)

View full text Add to dashboard Cite

Abstract-Localized spectro-temporal analysis is a novel feature extraction strategy in speech recognition, which was inspired by neurophysiological findings. Here we perform phone recognition experiments on features that are extracted from the patches of the critical-band log-energy spectrum by applying the two-dimensional cosine transform. We find that in phone recognition experiments the proposed feature set yields results similar to the standard MFCC features under clean conditions, while it provides a significantly smaller performance degradation in noisy conditions. Moreover, we show that the new and the standard features can be readily combined to improve the recognition accuracy still further.

show abstract

Section: Noisy Speech Experimentsmentioning

confidence: 80%

Section: Localized Spectro-temporal Featuresmentioning

confidence: 99%

Phone recognition experiments with 2D-DCT spectro-temporal features

Kovács

Tóth

2011

2011 6th IEEE International Symposium on Applied Computational Intelligence and Informatics (SACI)

View full text Add to dashboard Cite

show abstract

“…This implies that either the transition models were not considerably hurt by the exposure to the temporal randomness in the pseudo-utterances, or that the transition models have only a limited impact on the overall quality of an HMM. While the authors hypothesize that both factors are involved in some way, the latter hypothesis has been supported by studies that went as far as setting all HMM state transition probabilities to a constant, while attaining meaningful performance [e.g., Ketabdar and Bourlard (2010)]. Finally, it is noted the pseudo-utterances still retain a good portion of the meaningful temporal information through the first-and second-order time derivatives included in the frame feature vectors, and this information is learned by the state models of the HMM system.…”

Section: Temporal Smoothingmentioning

confidence: 98%

Deep neural network training for whispered speech recognition using small databases and generative model sampling

Ghaffarzadegan

Bořil

Hansen

2017

Int J Speech Technol

View full text Add to dashboard Cite

the original training data to train the DNN-HMM acoustic model of a speech recognizer. The proposed method is evaluated on small-sized sets of neutral and whisper data drawn from the UT-Vocal Effort II corpus. It is shown that phoneme error rates (PERs) of a DNN-HMM based speech recognizer are considerably reduced when incorporating the generated pseudo-samples in the training process, with + 79.0 and + 45.6% relative PER improvements for neutral-neutral training/test and whisper-whisper training/test scenarios, respectively.

show abstract

“…The origins of this technology go back to the era of shallow networks, which-just as DNNs-were trained on a block of consecutive input vectors. Some authors observed that the posterior estimates obtained can be "enhanced" by training yet another network-but this time on a sequence of output vectors coming from the first network [22]. Other authors refer to this approach as the "hierarchical modeling" [23][24][25] or the "stacked modeling" method [26].…”

Section: Introductionmentioning

confidence: 99%

Phone recognition with hierarchical convolutional deep maxout networks

Tóth

2015

J AUDIO SPEECH MUSIC PROC.

View full text Add to dashboard Cite

Deep convolutional neural networks (CNNs) have recently been shown to outperform fully connected deep neural networks (DNNs) both on low-resource and on large-scale speech tasks. Experiments indicate that convolutional networks can attain a 10-15 % relative improvement in the word error rate of large vocabulary recognition tasks over fully connected deep networks. Here, we explore some refinements to CNNs that have not been pursued by other authors. First, the CNN papers published up till now used sigmoid or rectified linear (ReLU) neurons. We will experiment with the maxout activation function proposed recently, which has been shown to outperform the rectifier activation function in fully connected DNNs. We will show that the pooling operation of CNNs and the maxout function are closely related, and so the two technologies can be readily combined to build convolutional maxout networks. Second, we propose to turn the CNN into a hierarchical model. The origins of this approach go back to the era of shallow nets, where the idea of stacking two networks on each other was relatively well known. We will extend this method by fusing the two networks into one joint deep model with many hidden layers and a special structure. We will show that with the hierarchical modelling approach, we can reduce the error rate of the network on an expanded context of input. In the experiments on the Texas Instruments Massachusetts Institute of Technology (TIMIT) phone recognition task, we find that a CNN built from maxout units yields a relative phone error rate reduction of about 4.3 % over ReLU CNNs. Applying the hierarchical modelling scheme to this CNN results in a further relative phone error rate reduction of 5.5 %. Using dropout training, the lowest error rate we get on TIMIT is 16.5 %, which is currently the best result. Besides experimenting on TIMIT, we also evaluate our best models on a low-resource large vocabulary task, and we find that all the proposed modelling improvements give consistently better results for this larger database as well.

show abstract

Enhanced Phone Posteriors for Improving Speech Recognition Systems

Cited by 34 publications

References 26 publications

Phone recognition experiments with 2D-DCT spectro-temporal features

Phone recognition experiments with 2D-DCT spectro-temporal features

Deep neural network training for whispered speech recognition using small databases and generative model sampling

Phone recognition with hierarchical convolutional deep maxout networks

Contact Info

Product

Resources

About