Filters








100 Hits in 9.0 sec

Phonet: A Tool Based on Gated Recurrent Neural Networks to Extract Phonological Posteriors from Speech

J.C. Vásquez-Correa, Philipp Klumpp, Juan Rafael Orozco-Arroyave, Elmar Nöth
2019 Interspeech 2019  
This paper presents a tool to extract phonological posteriors directly from speech signals.  ...  The proposed method consists of a bank of parallel bidirectional recurrent neural networks to estimate the posterior probabilities of the occurrence of different phonological classes.  ...  The authors also thank to CODI from University of Antioquia, grants # PRG2015-7683 and 2017-15530.  ... 
doi:10.21437/interspeech.2019-1405 dblp:conf/interspeech/Vasquez-CorreaK19 fatcat:moktwyx5ozczferhn57f7owk2q

DEEP DISCRIMINATIVE AND GENERATIVE MODELS FOR SPEECH PATTERN RECOGNITION [chapter]

Li Deng, Navdeep Jaitly
2015 Handbook of Pattern Recognition and Computer Vision  
We focus on speech recognition but our analysis is applicable to other domains.  ...  In this chapter we describe deep generative and discriminative models as they have been applied to speech recognition.  ...  The second major area where DNNs have made a significant impact in speech recognition is to move from hand-crafted features to automatic feature extraction from raw signals.  ... 
doi:10.1142/9789814656535_0002 fatcat:2ovjgqq4njgohffzvdnnut6si4

Automatic boost articulation therapy in adults with dysarthria: Acceptability, usability and user interaction

Viviana Mendoza Ramos, Juan C. Vasquez‐Correa, Rani Cremers, Leen Van Den Steen, Elmar Nöth, Marc De Bodt, Gwen Van Nuffelen
2021 International journal of language and communication disorders  
to benefit from speech therapy.  ...  The tool incorporates automated methods to detect phonological errors, which are specifically designed to analyse Dutch speech production.  ...  state-of-the-art gated recurrent neural networks (GRNN).  ... 
doi:10.1111/1460-6984.12647 fatcat:wcfw2thh3vbqfdnm3brm6aibim

Predicting speech from a cortical hierarchy of event-based time scales

Lea-Maria Schmitt, Julia Erb, Sarah Tune, Anna U. Rysop, Gesa Hartwigsen, Jonas Obleser
2021 Science Advances  
Heinrich for discussions on natural language processing; and M. Wöstmann for suggestions to the reading task design.  ...  Acknowledgments: We thank A. Herrmann, C. Mergner, M. Naujokat, A. Ruhe, and S. Meyn for help with data acquisition; C. Sickert for help in preparing the text corpus; M.  ...  Modeling neural speech prediction with artificial neural networks. (A) Bottom: Participants listened to a story (gray waveform) during fMRI.  ... 
doi:10.1126/sciadv.abi6070 pmid:34860554 pmcid:PMC8641937 fatcat:fvk6qwmkw5bypcnhvwc6gs7gci

Unsupervised Learning for Expressive Speech Synthesis

Igor Jauk
2018 IberSPEECH 2018  
Finally, accounting for the new tendencies in the speech synthesis world, deep neural network based expressive speech synthesis is designed and tested.  ...  Nowadays, especially with the upswing of neural networks, speech synthesis is almost totally data driven.  ...  Acknowledgements First of all I would like to thank Antonio Bonafonte for his help, lead and patience, and for the opportunity to work and to develop this work in his group.  ... 
doi:10.21437/iberspeech.2018-38 dblp:conf/iberspeech/Jauk18 fatcat:6zogjdy3gjgslfbbgrqirjzsx4

Deep Learning: Methods and Applications

Li Deng
2014 Foundations and Trends® in Signal Processing  
[235] applied a deep recurrent auto encoder neural network to remove noise in the input features for robust speech recognition.  ...  The HMM, based on dynamic programing operations, is a convenient tool to help port the strength of a static classifier to handle dynamic or sequential patterns.  ... 
doi:10.1561/2000000039 fatcat:vucffxhse5gfhgvt5zphgshjy4

Bangla Natural Language Processing: A Comprehensive Analysis of Classical, Machine Learning, and Deep Learning Based Methods

Ovishake Sen, Mohtasim Fuad, Md. Nazrul Islam, Jakaria Rabbi, Mehedi Masud, Md. Kamrul Hasan, Md. Abdul Awal, Awal Ahmed Fime, Md. Tahmid Hasan Fuad, Delowar Sikder, Md. Akil Raihan Iftee
2022 IEEE Access  
The studies are mainly concentrated on the specific domains of BNLP, such as sentiment analysis, speech recognition, optical character recognition, and text summarization.  ...  There is an apparent scarcity of resources that contain a comprehensive review of the recent BNLP tools and methods.  ...  The authors used a recurrent neural network for their model, and they have used a gated recurrent unit in their model for the recurrent neural network.  ... 
doi:10.1109/access.2022.3165563 fatcat:rmersduz6vbyjjczvobrebskmi

Bangla Natural Language Processing: A Comprehensive Review of Classical, Machine Learning, and Deep Learning Based Methods [article]

Ovishake Sen, Mohtasim Fuad, MD. Nazrul Islam, Jakaria Rabbi, MD. Kamrul Hasan, Mohammed Baz, Mehedi Masud, Md. Abdul Awal, Awal Ahmed Fime, Md. Tahmid Hasan Fuad, Delowar Sikder, MD. Akil Raihan Iftee
2021 arXiv   pre-print
There is an apparent scarcity of resources that contain a comprehensive study of the recent BNLP tools and methods.  ...  The studies are mainly concentrated on the specific domains of BNLP, such as sentiment analysis, speech recognition, optical character recognition, and text summarization.  ...  Acknowledgment: The authors would like to thank for the support from Taif University Researchers Supporting Project number (TURSP-2020/239), Taif University, Taif, Saudi Arabia.  ... 
arXiv:2105.14875v2 fatcat:kvqmgxpthvh2fj7jza64n6kaiq

NUVA: A Naming Utterance Verifier for Aphasia Treatment

David S. Barbera, Mark Huckvale, Victoria Fleming, Emily Upton, Henry Coley-Fisher, Catherine Doogan, Ian Shaw, William Latham, Alexander P. Leff, Jenny Crinion
2021 Computer Speech and Language  
This performance was not only significantly better than a baseline created for this study using one of the leading commercially available ASRs (Google speech-to-text service) but also comparable in some  ...  When tested on eight native British-English speaking PWA the system's performance accuracy ranged between 83.6% to 93.6%, with a 10-fold cross-validation mean of 89.5%.  ...  To generate posteriorgrams, our system NUVA replaces their Gaussian Mixture Model trained on unlabelled speech corpora with an acoustic model to yield phone-based posteriorgrams using a deep neural network  ... 
doi:10.1016/j.csl.2021.101221 pmid:34483474 pmcid:PMC8117974 fatcat:thrahglhb5dczojt4vprpnucwu

NUVA: A Naming Utterance Verifier for Aphasia Treatment [article]

David Sabate Barbera, Mark Huckvale, Victoria Fleming, Emily Upton, Henry Coley-Fisher, Catherine Doogan, Ian Shaw, William Latham, Alexander P. Leff, Jenny Crinion
2021 arXiv   pre-print
This performance was not only significantly better than a baseline created for this study using one of the leading commercially available ASRs (Google speech-to-text service) but also comparable in some  ...  When tested on eight native British-English speaking PWA the system's performance accuracy ranged between 83.6% to 93.6%, with a 10-fold cross-validation mean of 89.5%.  ...  fed to a Deep Neural Network (3) which outputs a vector of posterior probabilities or posteriorgram (4).  ... 
arXiv:2102.05408v1 fatcat:gvkxmcde6fbrbfnxmknq4p3zz4

Neural blackboard architectures of combinatorial structures in cognition

Frank van der Velde, Marc de Kamps
2006 Behavioral and Brain Sciences  
This paper aims to show that these problems can be solved by means of neural "blackboard" architectures. For this purpose, a neural blackboard architecture for sentence structure is presented.  ...  In his recent book on the foundations of language, Jackendoff described four fundamental problems for a neural instantiation of combinatorial structures: the massiveness of the binding problem, the problem  ...  John and McClelland (1990) presented a more flexible model based on a recurrent network.  ... 
doi:10.1017/s0140525x06009022 pmid:16542539 fatcat:yjti7u4gxnaojk6m6gspdasmka

Sentence Comprehension [chapter]

2014 Cognitive Neuroscience of Language  
Finally, the fifth lobe, which is hidden from view, is called the insula, a name based on the Latin word for "island."  ...  of Broca's area, a region known to contribute to phonological processing (see Chapters 5 and 6).  ... 
doi:10.4324/9781315764061-24 fatcat:ngyht6cvpffr3dsjeg24thesoy

High-frequency neural activity and human cognition: Past, present and possible future of intracranial EEG research

Jean-Philippe Lachaux, Nikolai Axmacher, Florian Mormann, Eric Halgren, Nathan E. Crone
2012 Progress in Neurobiology  
based on timing only.  ...  HFA was characterized by a posterior-anterior spread between 200 and 400 ms from the left temporal to the left frontal lobe.  ...  Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.pneurobio.2012. 06.008.  ... 
doi:10.1016/j.pneurobio.2012.06.008 pmid:22750156 pmcid:PMC3980670 fatcat:teaxt5dllfcmha56fhymhjmatm

Targets for a Comparative Neurobiology of Language

Justin T. Kiggins, Jordan A. Comins, Timothy Q. Gentner
2012 Frontiers in Evolutionary Neuroscience  
One longstanding impediment to progress in understanding the neural basis of language is the development of model systems that retain language-relevant cognitive behaviors yet permit invasive cellular  ...  It remains unknown, however, what a neuroscience of language perception may look like when instantiated at the cellular or network level.  ...  To test this idea, Maye et al. (2002) exposed 6-to 8-month-old infants to speech sounds varied along a phonetic continuum.  ... 
doi:10.3389/fnevo.2012.00006 pmid:22509163 pmcid:PMC3321487 fatcat:46oaibzahjeunmmipdxfh36xry

Computational Morphology with Neural Network Approaches [article]

Ling Liu
2021 arXiv   pre-print
This paper starts with a brief introduction to computational morphology, followed by a review of recent work on computational morphology with neural network approaches, to provide an overview of the area  ...  Neural network approaches have been applied to computational morphology with great success, improving the performance of most tasks by a large margin and providing new perspectives for modeling.  ...  neural networks (CNNs), recurrent neural networks (RNNs) and the Transformer.  ... 
arXiv:2105.09404v1 fatcat:6w4n7yjaevh6fpnauntzlfe64u
« Previous Showing results 1 — 15 out of 100 results