1,064 Hits in 5.5 sec

Multilingual acoustic models using distributed deep neural networks

G. Heigold, V. Vanhoucke, A. Senior, P. Nguyen, M. Ranzato, M. Devin, J. Dean
2013 2013 IEEE International Conference on Acoustics, Speech and Signal Processing  
Neural networks lend themselves naturally to parameter sharing across languages, and distributed implementations have made it feasible to train large networks.  ...  In this context, it is of paramount importance to train accurate acoustic models for many languages within given resource constraints such as data, processing power, and time.  ...  SUMMARY We presented an empirical comparison of mono-, cross-, and multilingual acoustic model training using deep neural networks.  ... 
doi:10.1109/icassp.2013.6639348 dblp:conf/icassp/HeigoldVSNRDD13 fatcat:ejxpbrj77jacjlhpelvbnnnaci

Multilingual exemplar-based acoustic model for the NIST Open KWS 2015 evaluation

Van Hai Do, Xiong Xiao, Haihua Xu, Eng Siong Chng, Haizhou Li
2015 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA)  
Specifically, kerneldensity model is used to replace GMM in HMM/GMM (Hidden Markov Model / Gaussian Mixture Model) or DNN in HMM/DNN (Hidden Markov Model / Deep Neural Network) acoustic model to predict  ...  In this paper, we investigate the use of the proposed non-parametric exemplar-based acoustic modeling for the NIST Open Keyword Search 2015 Evaluation.  ...  It is well known that using discriminative models e.g., multilayer perceptron (MLP), deep neural network (DNN) [36, 37] or discriminative training criteria [38] can significantly improve performance  ... 
doi:10.1109/apsipa.2015.7415338 dblp:conf/apsipa/DoXXCL15 fatcat:6sthbmtvvffftgs456xt7zm2ce

Multilingual deep neural network based acoustic modeling for rapid language adaptation

Ngoc Thang Vu, David Imseng, Daniel Povey, Petr Motlicek, Tanja Schultz, Herve Bourlard
2014 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
This paper presents a study on multilingual deep neural network (DNN) based acoustic modeling and its application to new languages.  ...  Using ten different languages from the Globalphone database, our studies reveal that crosslingual acoustic model transfer through multilingual DNNs is superior to unsupervised RBM pre-training and greedy  ...  MULTILINGUAL DEEP NEURAL NETWORKS For our studies, we use multilingual DNNs.  ... 
doi:10.1109/icassp.2014.6855086 dblp:conf/icassp/VuIPMSB14 fatcat:pxh4jbwbbjetplhdtcgr5hrb6i

Using generalized maxout networks and phoneme mapping for low resource ASR- a case study on Flemish-Afrikaans

Reza Sahraeian, Dirk Van Compernolle, Febe de Wet
2015 2015 Pattern Recognition Association of South Africa and Robotics and Mechatronics International Conference (PRASA-RobMech)  
Recently, multilingual deep neural networks (DNNs) have been successfully used to improve under-resourced speech recognizers.  ...  Index Terms-Low resource ASR, phoneme mapping, Kullback Leibler Divergence, multilingual deep neural network.  ...  CONCLUSION This paper presented an investigation of using generalized maxout networks and phoneme mappings for multilingual DNN based acoustic modeling.  ... 
doi:10.1109/robomech.2015.7359508 fatcat:vyl3l5zn6ne4fmkhi4tkb3embm

New types of deep neural network learning for speech recognition and related applications: an overview

Li Deng, Geoffrey Hinton, Brian Kingsbury
2013 2013 IEEE International Conference on Acoustics, Speech and Signal Processing  
We also describe the historical context in which acoustic models based on deep neural networks have been developed.  ...  of leveraging multiple languages or dialects that are more easily achieved with deep neural networks than with Gaussian mixture models.  ...  previous attempts to use neural networks for acoustic modeling.  ... 
doi:10.1109/icassp.2013.6639344 dblp:conf/icassp/DengHK13 fatcat:ct2qjsg5v5cp7dchtvdagabltm

Multilingual shifting deep bottleneck features for low-resource ASR

Quoc Bao Nguyen, Jonas Gehring, Markus Muller, Sebastian Stuker, Alex Waibel
2014 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
In this work, we propose a deep bottleneck feature architecture that is able to leverage data from multiple languages. We also show that tonal features are helpful for non-tonal languages.  ...  Furthermore, Gehring et al. proposed several multilingual deep neural network architectures with the connection of bottleneck feature extraction and acoustic model networks in order to create significantly  ...  The use of multilingual models in acoustic modeling is especially of use when The authors would like to thank Joshua Winebarger for his assistance in proofreading this paper.  ... 
doi:10.1109/icassp.2014.6854676 dblp:conf/icassp/NguyenGMSW14 fatcat:ufeirivtrffprayqkk6x7cvxhe

Very deep multilingual convolutional neural networks for LVCSR

Tom Sercu, Christian Puhrsch, Brian Kingsbury, Yann LeCun
2016 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
My colleagues at IBM The torch developers for an amazing deep learning tool Christian Szegedy for the figure of slide 3 The IARPA Babel program This effort uses the very limited language packs from IARPA  ...  Government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright annotation thereon.  ...  FC TOK CEB KAZ TEL LIT FC FC FC FC FC FC input c 2016 IBM Corporation Overview Very deep convolutional networks Small 3 × 3 kernels Multilingual training Shared convolutional  ... 
doi:10.1109/icassp.2016.7472620 dblp:conf/icassp/SercuPKL16 fatcat:lbld74rqtzb7dorhd2oq4ovfai

An Exploration towards Joint Acoustic Modeling for Indian Languages: IIIT-H Submission for Low Resource Speech Recognition Challenge for Indian Languages, INTERSPEECH 2018

Hari Krishna, Krishna Gurugubelli, Vishnu Vidyadhara Raju V, Anil Kumar Vuppala
2018 Interspeech 2018  
To exploit this property, a joint acoustic model has been trained for developing multilingual ASR system using a common phone-set.  ...  Sub-space Gaussian mixture models (SGMM), and recurrent neural networks (RNN) trained with connectionst temporal classification (CTC) objective function are explored for training joint acoustic models.  ...  Advancements in deep neural networks have greatly influenced the performances of speech recognition systems.  ... 
doi:10.21437/interspeech.2018-1584 dblp:conf/interspeech/VydanaGVV18 fatcat:3mrawh3kbndfzpdquu4hmt6j2e

Acoustic Modeling in Speech Recognition: A Systematic Review

Shobha Bhatt, Anurag Jain, Amita Dev
2020 International Journal of Advanced Computer Science and Applications  
Main issues in acoustic modeling such as feature extraction techniques, acoustic modeling units, speech corpora, classification methods, different tools used, language issues applied, and evaluation parameters  ...  The result of this review can be used to build a better speech recognition system by choosing a suitable acoustic modeling construct in SR.  ...  ANNs), deep neural networks(DNNs), and sequence to sequence acoustic modeling.  ... 
doi:10.14569/ijacsa.2020.0110455 fatcat:ezvwvhungrbv7eor4h34jpo2nu

Advances in Low Resource ASR: A Deep Learning Perspective

Hardik Sailor, Ankur Patil, Hemant Patil
2018 The 6th Intl. Workshop on Spoken Language Technologies for Under-Resourced Languages  
Recently, data augmentation, multilingual and cross-lingual approaches, transfer learning, etc. enable training deep learning architectures.  ...  This paper could be a good motivation for the researchers interested to work towards low resource ASR using deep learning techniques.  ...  However, to train deep architectures, such as Deep Neural Networks (DNN), a large amount of training data is required.  ... 
doi:10.21437/sltu.2018-4 dblp:conf/sltu/SailorPP18 fatcat:cnj5kkrxvnfd3dck46louwv7ya

Landmark-based consonant voicing detection on multilingual corpora

Xiang Kong, Xuesong Yang, Mark Hasegawa-Johnson, Jeung-Yoon Choi, Stefanie Shattuck-Hufnagel
2017 Journal of the Acoustical Society of America  
and(3) acoustic features computed using a convolutional neural network (CNN).  ...  All detectors are trained on English data (TIMIT),and tested on English, Turkish, and Spanish (performance measured using F1 and accuracy).  ...  learned from deep neural networks.  ... 
doi:10.1121/1.4987203 fatcat:mxqoowa55bfzlf65qsblei5ala

Multilingual Data Selection for Low Resource Speech Recognition

Samuel Thomas, Kartik Audhkhasi, Jia Cui, Brian Kingsbury, Bhuvana Ramabhadran
2016 Interspeech 2016  
Feature representations extracted from deep neural networkbased multilingual frontends provide significant improvements to speech recognition systems in low resource settings.  ...  The proposed multilingual features provide up to 15% relative improvement over baseline acoustic features on the VLLP languages.  ...  With deep neural networks (DNNs) becoming popular for acoustic modeling, several variants of these networks have been proposed for speech recognition in low resource settings [3] [4] [5] [6] [7] [8] [  ... 
doi:10.21437/interspeech.2016-598 dblp:conf/interspeech/ThomasACKR16 fatcat:rnumqswls5ek3mfbrscx5omjwy

Multilingual Speech Evaluation: Case Studies on English, Malay and Tamil [article]

Huayun Zhang, Ke Shi, Nancy F. Chen
2021 arXiv   pre-print
With the development of deep neural network (DNN), GOP was further optimized to predict better phone segmentation and posterior estimation [7, 8, 9] .  ...  Optimized on various ASR tasks, this model is designed as a combination of different neural networks: At the bottom levels, a stack of specially designed convolution neural network (CNN) running on 2D  ... 
arXiv:2107.03675v1 fatcat:mfjujvfd55eark4rth5ymxrpw4

Acoustic Modeling Based on Deep Learning for Low-Resource Speech Recognition: An Overview

Chongchong Yu, Meng Kang, Yunbing Chen, Jiajia Wu, Xia Zhao
2020 IEEE Access  
ACOUSTIC MODELS WITH NEURAL NETWORKS In the current speech recognition system based on neural networks, the common hybrid DNN acoustic model has been gradually replaced by more accurate RNN or CNN.  ...  It is well known that the depth of neural networks is critical for acoustic modeling.  ... 
doi:10.1109/access.2020.3020421 fatcat:uiws6fazpnghzj5lmkkmy7ol3y

Multi-Spectral Widefield Microscopy of the Beating Heart Through Post-Acquisition Synchronization and Unmixing

Christian Jaques, Linda Bapst-Wicht, Daniel F. Schorderet, Michael Liebling
2019 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019)  
State-of-the-art acoustic models for Automatic Speech Recognition (ASR) are based on Hidden Markov Models (HMM) and Deep Neural Networks (DNN) and often require thousands of hours of transcribed speech  ...  The goal of this thesis is to improve current state-of-the-art acoustic modeling techniques in general for ASR, with a particular focus on multilingual ASR and cross-lingual adaptation.  ...  by a deep neural network model.  ... 
doi:10.1109/isbi.2019.8759472 dblp:conf/isbi/JaquesBSL19 fatcat:flypznnglbfrzm3ayf6tsfof34
« Previous Showing results 1 — 15 out of 1,064 results