Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
SP, IPSJ-SLP, EA, SIP [detail] |
2023-03-01 09:30 |
Okinawa |
(Primary: On-site, Secondary: Online) |
A Study on Scheduled Sampling for Neural Transducer-based ASR Takafumi Moriya, Takanori Ashihara, Hiroshi Sato, Kohei Matsuura, Tomohiro Tanaka, Ryo Masumura (NTT) EA2022-100 SIP2022-144 SP2022-64 |
In this paper, we propose scheduled sampling approaches suited for the recurrent neural network-transducer (RNNT) that i... [more] |
EA2022-100 SIP2022-144 SP2022-64 pp.147-152 |
EA, SIP, SP, IPSJ-SLP [detail] |
2022-03-02 10:20 |
Okinawa |
(Primary: On-site, Secondary: Online) |
A Study on Hybrid RNN-T/Attention-based Streaming ASR with Triggered Chunkwise Attention and Dual Internal Language Model Integration Takafumi Moriya, Takanori Ashihara, Atsushi Ando, Hiroshi Sato, Tomohiro Tanaka, Kohei Matsuura, Ryo Masumura, Marc Delcroix (NTT), Takahiro Shinozaki (Tokyo Tech) EA2021-78 SIP2021-105 SP2021-63 |
In this paper we propose improvements to our recently proposed hybrid RNN-T/Attention architecture that includes a share... [more] |
EA2021-78 SIP2021-105 SP2021-63 pp.90-95 |
SP, EA, SIP |
2020-03-02 13:00 |
Okinawa |
Okinawa Industry Support Center (Cancelled but technical report was issued) |
Data augmentation for ASR system by using locally time-reversed speech
-- Temporal inversion of feature sequence -- Takanori Ashihara, Tomohiro Tanaka, Takafumi Moriya, Ryo Masumura, Yusuke Shinohara, Makio Kashino (NTT) EA2019-110 SIP2019-112 SP2019-59 |
Data augmentation is one of the techniques to mitigate overfitting and improve robustness against several acoustic varia... [more] |
EA2019-110 SIP2019-112 SP2019-59 pp.53-58 |
SP, EA, SIP |
2020-03-02 15:45 |
Okinawa |
Okinawa Industry Support Center (Cancelled but technical report was issued) |
Performance evaluation of distilling knowledge using encoder-decoder for CTC-based automatic speech recognition systems Takafumi Moriya, Hiroshi Sato, Tomohiro Tanaka, Takanori Ashihara, Ryo Masumura, Yusuke Shinohara (NTT) EA2019-131 SIP2019-133 SP2019-80 |
We present a novel training approach for connectionist temporal classification (CTC) -based automatic speech recognition... [more] |
EA2019-131 SIP2019-133 SP2019-80 pp.175-180 |
WIT, SP |
2019-10-26 17:00 |
Kagoshima |
Daiichi Institute of Technology |
Neural Whispered Speech Detection with Imbalanced Learning Takanori Ashihara, Yusuke Shinohara, Hiroshi Sato, Takafumi Moriya, Kiyoaki Matsui, Yoshikazu Yamaguchi (NTT) SP2019-26 WIT2019-25 |
In this paper, we present a neural whispered-speech detection technique that offers utterance-level classification of wh... [more] |
SP2019-26 WIT2019-25 pp.51-56 |
SP |
2018-08-27 11:35 |
Kyoto |
Kyoto Univ. |
SP2018-23 |
(To be available after the conference date) [more] |
SP2018-23 pp.7-8 |
SIP, EA, SP, MI (Joint) [detail] |
2018-03-19 13:00 |
Okinawa |
|
[Poster Presentation]
Investigation about LSTM Post-filter for Voice Activity Detection Kiyoaki Matsui, Takafumi Moriya, Takaaki Fukutomi, Yusuke Shinohara, Yoshikazu Yamaguchi, Manabu Okamoto, Yushi Aono (NTT) EA2017-109 SIP2017-118 SP2017-92 |
[more] |
EA2017-109 SIP2017-118 SP2017-92 pp.45-50 |
SP, SIP, EA |
2017-03-02 09:00 |
Okinawa |
Okinawa Industry Support Center |
[Poster Presentation]
Study of branch selecting DNN acoustic model for robustness to environmental variation Takafumi Moriya, Taichi Asami, Yoshikazu Yamaguchi, Yushi Aono (NTT) EA2016-131 SIP2016-186 SP2016-126 |
The performance of speech recognition tasks can be significantly improved by the use of deep neural networks (DNN). Spee... [more] |
EA2016-131 SIP2016-186 SP2016-126 pp.277-282 |
NLC, IPSJ-NL, SP, IPSJ-SLP (Joint) [detail] |
2015-12-02 13:55 |
Aichi |
Nagoya Inst of Tech. |
Automation of high performance system building for large vocabulary speech recognition using evolution strategy with pareto optimality Takafumi Moriya, Tomohiro Tanaka, Takahiro Shinozaki (Tokyo Tech), Shinji Watanabe (MERL), Kevin Duh (NAIST) SP2015-75 |
The performance of speech recognition tasks can be significantly improved by the use of deep neural networks (DNN). Howe... [more] |
SP2015-75 pp.31-36 |