Zobrazeno 1 - 10
of 22
pro vyhledávání: '"Murali Karthick Baskar"'
Publikováno v:
Interspeech 2022.
Masked speech modeling (MSM) methods such as wav2vec2 or w2v-BERT learn representations over speech frames which are randomly masked within an utterance. While these methods improve performance of Automatic Speech Recognition (ASR) systems, they have
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::950aa5ea6024ec86da39dbb74c4e24d6
Autor:
Shinji Watanabe, Jan Cernocky, Ramón Fernandez Astudillo, Lukas Burget, Murali Karthick Baskar
Publikováno v:
ICASSP
Self-supervised ASR-TTS models suffer in out-of-domain data conditions. Here we propose an enhanced ASR-TTS (EAT) model that incorporates two main features: 1) The ASR$\rightarrow$TTS direction is equipped with a language model reward to penalize the
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::59ad37018c8778608b4237075aa261f2
http://arxiv.org/abs/2104.07474
http://arxiv.org/abs/2104.07474
Autor:
Shinji Watanabe, Yuya Fujita, Murali Karthick Baskar, Toru Taniguchi, Dung Tran, Aswin Shanmugam Subramanian, Xiaofei Wang
Publikováno v:
WASPAA
Speech enhancement systems, which denoise and dereverberate distorted signals, are usually optimized based on signal reconstruction objectives including the maximum likelihood and minimum mean square error. However, emergent end-to-end neural methods
Publikováno v:
Speech Communication. 92:64-76
In this paper, we propose using deep neural networks (DNN) as a regression model to estimate speaker-normalized features from un-normalized features. We consider three types of speaker-specific feature normalization techniques, viz., feature-space ma
Autor:
Takaaki Hori, Murali Karthick Baskar, Ramón Fernandez Astudillo, Shinji Watanabe, Jan Cernocký, Lukas Burget
Publikováno v:
INTERSPEECH
Sequence-to-sequence automatic speech recognition (ASR) models require large quantities of data to attain high performance. For this reason, there has been a recent surge in interest for unsupervised and semi-supervised training in such models. This
Autor:
Najim Dehak, Hirofumi Inaguma, Takaaki Hori, Murali Karthick Baskar, Shinji Watanabe, Jesús Villalba, Jaejin Cho
Publikováno v:
ICASSP
In this paper, we explore several new schemes to train a seq2seq model to integrate a pre-trained LM. Our proposed fusion methods focus on the memory cell state and the hidden state in the seq2seq decoder long short-term memory (LSTM), and the memory
Publikováno v:
ICASSP
This work explores better adaptation methods to low-resource languages using an external language model (LM) under the framework of transfer learning. We first build a language-independent ASR system in a unified sequence-to-sequence (S2S) architectu
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::fb4ef0b208410575406d6736f59dde8a
http://arxiv.org/abs/1811.02134
http://arxiv.org/abs/1811.02134
Autor:
Bhargav Pulugundla, Jan Cernocký, Santosh Kesiraju, Murali Karthick Baskar, Ekaterina Egorova, Lukas Burget, Martin Karafiat
Publikováno v:
INTERSPEECH