1 |
Using Automatic Speech Recognition to Optimize Hearing-Aid Time Constants
|
|
|
|
In: ISSN: 1662-4548 ; EISSN: 1662-453X ; Frontiers in Neuroscience ; https://hal.archives-ouvertes.fr/hal-03627441 ; Frontiers in Neuroscience, Frontiers, 2022, 16 (779062), ⟨10.3389/fnins.2022.779062⟩ ; https://www.frontiersin.org/articles/10.3389/fnins.2022.779062/full (2022)
|
|
BASE
|
|
Show details
|
|
2 |
Einfluss von Sprechtempo und Störgeräusch auf das Sprachverstehen im Göttinger und im HSM-Satztest ... : Impact of speech rate and noise on speech recognition in Göttingen and HSM sentence test ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
Einfluss von Sprechtempo und Störgeräusch auf das Sprachverstehen im Göttinger und im HSM-Satztest ... : Impact of speech rate and noise on speech recognition in Göttingen and HSM sentence test ...
|
|
|
|
BASE
|
|
Show details
|
|
4 |
Assessing the intelligibility and acoustic changes of time-processed speech
|
|
|
|
In: http://rave.ohiolink.edu/etdc/view?acc_num=case1586637814204979 (2020)
|
|
BASE
|
|
Show details
|
|
5 |
ФУНКЦИОНИРОВАНИЕ КОМПРЕССИИ В УСТНОМ ПЕРЕВОДЕ НА ПРЕСС-КОНФЕРЕНЦИЯХ ... : Functioning of compression in interpretation at press conferences ...
|
|
|
|
BASE
|
|
Show details
|
|
6 |
Speech recognition model compression
|
|
|
|
Abstract:
Speech recognition models are widely deployed in mobile and embedded devices. However, the base architecture with which these models are developed is usually made of neural networks with bigger size and millions of model parameters. In this report, we investigate three compression schemes for these neural network architecture with a trade-off on accuracy and compressed model size. Also, we perform sensitivity analysis on the network parameters with known perturbations to determine the best compression scheme for a particular layer. The first compression scheme deployed is k-means clustering. This helps in generating clusters which are used for weight sharing and hence reduction in the total number of parameters required. Secondly, we employ svd based compression on various network layer parameters and achieve the best compression using svd in the case of a large vocabulary continuous speech recognition model. Finally, a two-stage compression scheme using k-means and Huffman coding is proposed. We have investigated these compression schemes on keyword spotter speech recognition system and the Baidu’s DeepSpeech large vocabulary continuous speech recognition model and have shown 58.3% reduction in size for only a 3.4% drop in accuracy and 45% reduction in size for only a 1.21% drop in accuracy respectively. ; Electrical and Computer Engineering
|
|
Keyword:
Clustering; K-means; Model compression; Speech recognition; Svd
|
|
URL: https://doi.org/10.26153/tsw/5437 https://hdl.handle.net/2152/78350
|
|
BASE
|
|
Hide details
|
|
7 |
Dealing with linguistic mismatches for automatic speech recognition
|
|
|
|
BASE
|
|
Show details
|
|
8 |
A First Summarization System of a Video in a Target Language
|
|
|
|
In: MISSI 2018 - 11th edition of the International Conference on Multimedia and Network Information Systems ; https://hal.archives-ouvertes.fr/hal-01819720 ; MISSI 2018 - 11th edition of the International Conference on Multimedia and Network Information Systems, Sep 2018, Wrocław, Poland. pp.1-12 (2018)
|
|
BASE
|
|
Show details
|
|
9 |
Manipulation of Auditory Feedback in Individuals with Normal Hearing and Hearing Loss
|
|
|
|
In: Electronic Thesis and Dissertation Repository (2017)
|
|
BASE
|
|
Show details
|
|
10 |
Speech referenced dynamic compression limiting: improving loudness comfort and acoustic safety
|
|
|
|
BASE
|
|
Show details
|
|
11 |
On The (Un)importance of Working Memory in Speech-in-Noise Processing for Listeners with Normal Hearing Thresholds
|
|
|
|
In: FRONTIERS IN PSYCHOLOGY , 7 (ARTN 126) (2016) (2016)
|
|
BASE
|
|
Show details
|
|
13 |
Voice For The Mute
|
|
|
|
In: Research outputs 2014 to 2021 (2015)
|
|
BASE
|
|
Show details
|
|
14 |
A Patient-Centered, Provider-Facilitated Approach to the Refinement of Nonlinear Frequency Compression Parameters Based on Subjective Preference Ratings of Amplified Sound Quality
|
|
|
|
In: ETSU Faculty Works (2015)
|
|
BASE
|
|
Show details
|
|
15 |
Intelligibility of speech produced by children with hearing loss : conventional amplification versus nonlinear frequency compression in hearing aids
|
|
|
|
BASE
|
|
Show details
|
|
16 |
Effect of Compression Ratio on Perception of Time Compressed Phonemically Balanced Words in Kannada and Monosyllables
|
|
|
|
In: Audiology Research; Volume 5; Issue 1; Pages: 128 (2015)
|
|
BASE
|
|
Show details
|
|
17 |
ЕЩЕ РАЗ О ЗАКОНЕ ЭКОНОМИИ В ПОВСЕДНЕВНОЙ СПОНТАННОЙ РЕЧИ
|
|
БОГДАНОВА-БЕГЛАРЯН НАТАЛЬЯ ВИКТОРОВНА. - : Федеральное государственное бюджетное образовательное учреждение высшего профессионального образования «Омский государственный университет им. Ф.М. Достоевского», 2014
|
|
BASE
|
|
Show details
|
|
19 |
Aided cortical response, speech intelligibility, consonant perception and functional performance of young children using conventional amplification or nonlinear frequency compression
|
|
|
|
BASE
|
|
Show details
|
|
|
|