Zobrazeno 1 - 10
of 10
pro vyhledávání: '"Minguang Song"'
Autor:
Minguang Song, Yunxin Zhao
Publikováno v:
ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::a60dd92f6aa62bf5ddf3ad8110bfc2cb
https://doi.org/10.1109/icassp43922.2022.9747525
https://doi.org/10.1109/icassp43922.2022.9747525
Publikováno v:
BHI
Amyotrophic lateral sclerosis (ALS) patients experience progressive speech deterioration due to muscle paralysis, leading to eventual loss of verbal communication capability. Text-to-speech synthesis (TTS) is an important technology for speech genera
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::0c84769e145e1d9f9d62077945033e31
https://doi.org/10.1109/bhi50953.2021.9508522
https://doi.org/10.1109/bhi50953.2021.9508522
Publikováno v:
SLT
Label smoothing has been shown as an effective regularization approach for deep neural networks. Recently, a context-sensitive label smoothing approach was proposed for training RNNLMs that improved word error rates on speech recognition tasks. Despi
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::9aa62345416798efd0e329ab797db426
https://doi.org/10.1109/slt48900.2021.9383598
https://doi.org/10.1109/slt48900.2021.9383598
Publikováno v:
ICASSP
Recurrent neural network language models (RNNLMs) have become very successful in many natural language processing tasks. However, RNNLMs trained with a cross entropy loss function and hard output targets are prone to overfitting, which weakens the la
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::2f8340660e6b55796539f1b07f00cd2d
https://doi.org/10.1109/icassp40776.2020.9053589
https://doi.org/10.1109/icassp40776.2020.9053589
Publikováno v:
IEEE/ACM transactions on audio, speech, and language processing. 26(12)
We investigate a structured sparse spectral transform method for voice conversion (VC) to perform frequency warping and spectral shaping simultaneously on high-dimensional (D) STRAIGHT spectra. Learning a large transform matrix for high-D data often
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::758dcfc861b20a81c409d545f9ac2918
https://pubmed.ncbi.nlm.nih.gov/31984214
https://pubmed.ncbi.nlm.nih.gov/31984214
Publikováno v:
IEEE J Biomed Health Inform
Amyotrophic lateral sclerosis (ALS) results in progressive paralysis of voluntary muscles throughout the body. As speech deteriorates, individuals rely on pre-programmed messages available on commercial speech generating devices to communicate using
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::762a784505a4b358d4d761fde9a7a7c3
https://pubmed.ncbi.nlm.nih.gov/31880571
https://pubmed.ncbi.nlm.nih.gov/31880571
Publikováno v:
SLT
The cross entropy (CE) loss function is commonly adopted for neural network language model (NNLM) training. Although this criterion is largely successful, as evidenced by the quick advance of NNLM, minimizing CE only maximizes likelihood of training
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::0194d7c22238142e88e8184dbf50ed93
https://doi.org/10.1109/slt.2018.8639649
https://doi.org/10.1109/slt.2018.8639649
Publikováno v:
ICASSP
We propose to exploit the potential of multiple word clusterings in class-based recurrent neural network (RNN) language models for ensemble RNN language modeling. By varying the clustering criteria and the space of word embedding, different word clus
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::e836964515e118ca1616165c20323bad
https://doi.org/10.1109/icassp.2017.7953255
https://doi.org/10.1109/icassp.2017.7953255
Autor:
Minguang Song, Ping Guo
Publikováno v:
Journal of Software. 9
Recently, spatial principal component analysis of census transform histograms (PACT) was proposed to recognize instance and categories of places or scenes in an image. An improved representation called Local Difference Binary Pattern (LDBP) also was
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::02a5696109d308bf6ea90cf31b68ef23
https://doi.org/10.4304/jsw.9.1.203-210
https://doi.org/10.4304/jsw.9.1.203-210
Autor:
Minguang Song, Ping Guo
Publikováno v:
CIS
Recently, spatial principal component analysis of census transform histograms (PACT) was proposed to recognize instance and categories of places or scenes in an image. When combining PACT with Local difference Magnitude Binary Pattern (LMBP), a new r
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::bf2ab2b7de60c180e00b7b6f019eb79f
https://doi.org/10.1109/cis.2012.84
https://doi.org/10.1109/cis.2012.84