Hierarchical Neural Network Structures for Phoneme Recognition

Vasquez, Daniel; Gruhn, Rainer; Minker, Wolfgang

doi:10.1007/978-3-642-34425-1

Cited by 7 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, there is no pair of separate English words which are identical except that where one member of the pair has breathy [p] where the other has non breathy [p]. (pp.81-82) Vasquez, Gruhn, and Minker (2012) explain "phonemes are realized as individual sounds (phones) i.e., phones are instances of phonemes. Each phoneme is characterized by a prototype which mainly describes the way phones are produced (articulation).…”

Section: Discussionmentioning

confidence: 99%

Phonemes assimilation in Ed Sheeran’s Songs

Yuliana¹,

Hadi²,

Panjaitan³

2020

ARDU

View full text Add to dashboard Cite

Background: Within phonemes as English spoken by the native speakers, it sometimes undergoes simplication to ease the native speakers in expressing their feelings. That is why, it is common for them to speak English in high speed along with their emotion. As the result they make a ‘shortcut’ to get ease of their pronunciation. Purpose: this research is meant to analyze phoneme assimilation in the three songs of Ed Sheeran. Design and methods: This research uses qualitative method with the content of analyzing the phonemes assimilation in the songs. Results: The findings revealed 35 kinds phonemes assimilation in the three songs of Ed Sheeran in Divide album. There are 29 regressive assimilations and 6 progressive assimilation

show abstract

Section: Discussionmentioning

confidence: 99%

Phonemes assimilation in Ed Sheeran’s Songs

Yuliana¹,

Hadi²,

Panjaitan³

2020

ARDU

View full text Add to dashboard Cite

show abstract

“…As a middle ground between assuming gold phonetic transcriptions (cf. Section 6) and no transcriptions at all, we use noisy transcriptions by running speech recognizers for other languages on the Spanish speech: Russian (ru), Hungarian (hu) and Czech (cz) (Vasquez et al, 2012). These distantly related languages were chosen to be a better approximation to the low-resource scenario.…”

Section: Alignment Evaluationmentioning

confidence: 99%

An Attentional Model for Speech Translation Without Transcription

Duong¹,

Anastasopoulos²,

Chiang³

et al. 2016

Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

128

116

View full text Add to dashboard Cite

For many low-resource languages, spoken language resources are more likely to be annotated with translations than transcriptions. This bilingual speech data can be used for word-spotting, spoken document retrieval, and even for documentation of endangered languages. We experiment with the neural, attentional model applied to this data. On phoneto-word alignment and translation reranking tasks, we achieve large improvements relative to several baselines. On the more challenging speech-to-word alignment task, our model nearly matches GIZA++'s performance on gold transcriptions, but without recourse to transcriptions or to a lexicon.

show abstract

“…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]. Two trivial improvements to this approach are when the upper net downsamples the output of the lower one [24,27] and/or when it uses the output of some bottleneck layer instead of the uppermost softmax layer [25,26].…”

Section: Introductionmentioning

confidence: 99%

“…Other authors refer to this approach as the "hierarchical modeling" [23][24][25] or the "stacked modeling" method [26]. Two trivial improvements to this approach are when the upper net downsamples the output of the lower one [24,27] and/or when it uses the output of some bottleneck layer instead of the uppermost softmax layer [25,26]. Veselý's proposal was to treat this hierarchical construct as one joint model, and he also explained why the compound structure can be interpreted as a deep convolutional network [21].…”

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

Hierarchical Neural Network Structures for Phoneme Recognition

Cited by 7 publications

References 0 publications

Phonemes assimilation in Ed Sheeran’s Songs

Phonemes assimilation in Ed Sheeran’s Songs

An Attentional Model for Speech Translation Without Transcription

Phone recognition with hierarchical convolutional deep maxout networks

Contact Info

Product

Resources

About