Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
PRMU, IBISML, IPSJ-CVIM |
2024-03-03 15:00 |
Hiroshima |
Hiroshima Univ. Higashi-Hiroshima campus (Primary: On-site, Secondary: Online) |
The Generalized Denoising Autoencoder with Tweedie's Formula Yuta Aishima (NAIST), Sho Sonoda (RIKEN), Noboru Isobe (Tokyo Univ.), Kazushi Ikeda (NAIST) IBISML2023-40 |
Denoising autoencoders learn the score of the data-generating distribution, i.e., $nabla log p(x)$. However, theoretical... [more] |
IBISML2023-40 pp.1-5 |
IBISML |
2020-10-21 15:35 |
Online |
Online |
IBISML2020-24 |
The goal of this study is to understand the information processing mechanism in a deep neural network (DNN) as a curve $... [more] |
IBISML2020-24 p.42 |
QIT (2nd) |
2019-11-19 11:30 |
Tokyo |
Gakushuin University |
Fast quantum algorithm for data approximation by optimized random features Hayata Yamasaki (UTokyo), Sathyawageeswar Subramanian (University of Cambridge), Sho Sonoda (Riken), Masato Koashi (UTokyo) |
[more] |
|
MI |
2019-01-22 14:35 |
Okinawa |
|
Research on macular analysis based on OCT angiography image analysis Shinnosuke Yamaoka, Shinya Fukumoto, Masayuki Kashima, Kiminori Sato, Mutsumi Watanabe, Hideki Shiihara, Syozo Sonoda, Taiji Sakamoto (Kagoshima Univ.) MI2018-81 |
The purpose of this study is to ascertain whether the complexity of the shape of macular region obtained from OCT angiog... [more] |
MI2018-81 pp.87-92 |
NC, IBISML, IPSJ-BIO, IPSJ-MPS [detail] |
2018-06-13 15:25 |
Okinawa |
Okinawa Institute of Science and Technology |
Current Dipole Localization from EEG with Birth-Death Process Keita Nakamura (Waseda Univ.), Sho Sonoda (RIKEN), Hideitsu Hino (ISM), Masahiro Kawasaki (Univ. of Tsukuba), Shotaro Akaho (AIST), Noboru Murata (Waseda Univ.) IBISML2018-10 |
We explore the EEG source localization problem as the estimation of current dipoles. We formulate the relation between c... [more] |
IBISML2018-10 pp.67-74 |
SIP, EA, SP, MI (Joint) [detail] |
2018-03-19 11:00 |
Okinawa |
|
Research on Automatic Fundus Disease Diagnostic Support based on OCT Angiography Image Analysis Shinnosuke Yamaoka, Shinya Fukumoto, Masayuki Kashima, Kiminori Sato, Mutsumi Watanabe, Hideki Shiihara, Shozo Sonoda, Taiji Sakamoto (Kagoshima Univ.) MI2017-68 |
Optical coherence tomography (OCT) can shoot fundus photography used fundus examination. the purpose of this study is to... [more] |
MI2017-68 pp.19-24 |
PRMU, IBISML, IPSJ-CVIM [detail] |
2017-09-16 14:05 |
Tokyo |
|
[Invited Talk]
Transportation aspect of deep neural network Sho Sonoda (Waseda Univ.) PRMU2017-60 IBISML2017-32 |
What is happening in deep neural networks? In this talk, we formulate them as transport maps. From the transportation vi... [more] |
PRMU2017-60 IBISML2017-32 pp.185-188 |
IBISML |
2017-03-06 16:30 |
Tokyo |
Tokyo Institute of Technology |
New Lerning Algorythm of Neural Network using Integral Representation and Kernel Herding Takuo Matsubara, Sho Sonoda, Noboru Murata (Waseda Univ.) IBISML2016-103 |
A new learning algorithm for neural networks that converges at $mathcal{O}(frac{1}{n})$ with respect to model complexity... [more] |
IBISML2016-103 pp.25-31 |
SP, IPSJ-SLP, NLC, IPSJ-NL (Joint) [detail] |
2016-12-20 11:20 |
Tokyo |
NTT Musashino R&D |
Speaker Recognition Based on Features through 1-Dimensional Convolutional Neural Network Shohei Sonoda, Yufu Kasahara, Masato Inoue (Waseda Univ) SP2016-52 |
Most of the speaker recognition methods utilize the voice features of the mel-frequency cepstrum coefficients (MFCCs) an... [more] |
SP2016-52 pp.17-21 |
IBISML |
2016-11-17 14:00 |
Kyoto |
Kyoto Univ. |
Transportation aspect of infinitely deep denoising autoencoder Sho Sonoda, Noboru Murata (Waseda Univ.) IBISML2016-88 |
The hidden layers of a deep neural network extract feature of the input data. We propose to analyze the feature map of a... [more] |
IBISML2016-88 pp.297-304 |
NC, NLP (Joint) |
2016-01-28 15:25 |
Fukuoka |
Kyushu Institute of Technology |
Validation of the Effects of Ensemble Learning for i-vector-based Speaker Identification
-- Bagging vs Random forest -- Shohei Sonoda, Masato Inoue (Waseda Univ) NC2015-58 |
Currently, most speaker identification methods have been performed by i-vectors which represent the features of unique s... [more] |
NC2015-58 pp.13-16 |
IBISML |
2015-11-26 15:00 |
Ibaraki |
Epochal Tsukuba |
[Poster Presentation]
CNN Brake Scene Recognition with LiDAR Tatsunosuke Shimada, Takuo Matsubara, Sho Sonoda, Noboru Murata (Waseda Univ.), Patricia Ortal, Shinpei Kato (Nagoya Univ.) IBISML2015-61 |
[more] |
IBISML2015-61 pp.61-67 |
PRMU, IBISML, IPSJ-CVIM [detail] |
2014-09-02 15:45 |
Ibaraki |
|
Sampling Learning Algorithm by Oracle Distribution Sho Sonoda, Noboru Murata (Waseda Univ.) PRMU2014-52 IBISML2014-33 |
A new sampling learning algorithm for neural networks is proposed. Based on the integral representation of neural networ... [more] |
PRMU2014-52 IBISML2014-33 pp.137-142 |
NC, IPSJ-BIO, IBISML, IPSJ-MPS [detail] |
2014-06-26 15:00 |
Okinawa |
Okinawa Institute of Science and Technology |
Current Dipole Localization from EEG by Multiple Particle Filters and Model Selection Yuki Kaneda, Sho Sonoda (Waseda Univ.), Hideitsu Hino (Univ. of Tsukuba), Noboru Murata (Waseda Univ.) NC2014-5 IBISML2014-5 |
In this study, the location, the moment and the number of ionic current modeled as dipoles are estimated from EEG data. ... [more] |
NC2014-5 IBISML2014-5 pp.91-96 |
ET |
2012-10-27 15:30 |
Nagano |
|
Student's Stumble Detection using Writing Information and Time-series Model Hiroki Asai, Akari Nozawa, Shogo Sonoda (Waseda Univ.), Hayato Yamana (Waseda Univ /) ET2012-49 |
(To be available after the conference date) [more] |
ET2012-49 pp.65-70 |