Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
SIP, SP, EA, IPSJ-SLP [detail] |
2024-03-01 10:40 |
Okinawa |
(Primary: On-site, Secondary: Online) |
An Investigation into Weighting Strategies for Model Averaging in Continual Learning for Automatic Speech Recognition Kentaro Shinayama, Hiroshi Sato, Tomoharu Iwata, Takeshi Mori, Taichi Asami (NTT) EA2023-105 SIP2023-152 SP2023-87 |
In recent years, the application scope of speech recognition AI has expanded, enabling the acquisition of diverse data d... [more] |
EA2023-105 SIP2023-152 SP2023-87 pp.262-267 |
PRMU, IBISML, IPSJ-CVIM [detail] |
2023-03-03 17:00 |
Hokkaido |
Future University Hakodate (Primary: On-site, Secondary: Online) |
Continual learning combining Replay and Parameter Isolation Methods Ryota Adachi, Toshikazu Wada (Wakayama Univ.) PRMU2022-121 IBISML2022-128 |
Deep Neural Networks (DNN) are used for various tasks such as image classification, but when a new task is trained on a ... [more] |
PRMU2022-121 IBISML2022-128 pp.335-340 |
NC, IBISML, IPSJ-BIO, IPSJ-MPS [detail] |
2022-06-27 14:50 |
Okinawa |
(Primary: On-site, Secondary: Online) |
AI knowledge forgetting based on the importance of model parameters Tomoya Yamashita, Masanori Yamada (NTT) NC2022-3 IBISML2022-3 |
The accuracy of Deep Learning has been greatly improved by the progress of research in computer technology and Deep Lear... [more] |
NC2022-3 IBISML2022-3 pp.14-19 |
IE, ITS, ITE-AIT, ITE-ME, ITE-MMS [detail] |
2022-02-21 13:00 |
Online |
Online |
Domain Incremental Leaning with Adaptive Loss Functions Takumi Kawashima (UTokyo), Go Irie, Daiki Ikami (NTT), Kiyoharu Aizawa (UTokyo) ITS2021-30 IE2021-39 |
During domain incremental learning of image classification task, the distribution of images continually change, and mode... [more] |
ITS2021-30 IE2021-39 pp.31-36 |
MBE, NC, NLP, CAS (Joint) [detail] |
2020-10-30 10:50 |
Online |
Online |
statistical mechanical analysis of catastrophic forgetting in continual learning with teacher and student networks Haruka Asanuma, Shiro Takagi, Yoshihiro Nagano, Yuki Yoshida (Tokyo Univ.), Yasuhiko Igarashi (Tsukuba Univ.), Masato Okada (Tokyo Univ.) NC2020-18 |
When single neural networks sequentially learns more than one task, catastrophic forgetting occurs except for the last t... [more] |
NC2020-18 pp.50-55 |
ISEC, SITE, ICSS, EMM, HWS, BioX, IPSJ-CSEC, IPSJ-SPT [detail] |
2019-07-23 13:35 |
Kochi |
Kochi University of Technology |
Proposal of Spam Filter by Applying Neural Network for Mitigating Catastrophic Forgetting Shuiti Kawahara, Lu Chen, Hiroyuki Inaba (KIT) ISEC2019-32 SITE2019-26 BioX2019-24 HWS2019-27 ICSS2019-30 EMM2019-35 |
The trend of spam mails is changing. Since, in recent years, spam mails sent from overseas mail serversare increasing, s... [more] |
ISEC2019-32 SITE2019-26 BioX2019-24 HWS2019-27 ICSS2019-30 EMM2019-35 pp.171-178 |
NC, MBE (Joint) |
2019-03-04 10:45 |
Tokyo |
University of Electro Communications |
Reduction of Catastrophic Forgetting using Pseudorehearsal and Importance of Weight Shunta Nakano, Motonobu Hattori (Univ. of Yamanashi) NC2018-47 |
When artificial neural networks learn new information additionally, they encounter serious forgetting of information lea... [more] |
NC2018-47 pp.19-24 |
MBE, NC (Joint) |
2015-12-19 14:50 |
Aichi |
Nagoya Institute of Technology |
An Associative Memory Model with Forgetting Process by Eliminating Weak Synapses Takuya Akamine, Koji Kurata (Univ Ryukyu) NC2015-50 |
In associative memory models overloading beyond the memory capacity causes catastrophic forgetting. In order to avoid it... [more] |
NC2015-50 pp.25-30 |
MBE, NC (Joint) |
2011-11-25 14:15 |
Miyagi |
ECEI Departments, Graduate School of Engineering, Tohoku University |
Optimization of the Sparsely Encoded Associative Memory Model with Unit Replacement by the Mutual Information Jun Tsuzurugi (Okayama Univ. Science), Ryota Miyata (Tokyo Tech), Koji Kurata (Univ. Ryukyu) NC2011-83 |
According to the previous researches, the associative memory model with unit replacement, in which a few old units are r... [more] |
NC2011-83 pp.65-70 |