Presentation | 2017-11-07 [Invited Talk] Researchs on high-speed and efficient Deep Learning technologies Takuya Fukagai, Koichi Shirahata, Yasumoto Tomita, Tetsutaro Hashimoto, Atsushi Ike, Masafumi Yamazaki, Akihiko Kasagi, Tsuguchika Tabaru, Liuan Wang, Song Wang, Li Sun, Jun Sun, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Fujitsu laboratories have been doing Research and Development on AI technologies called "Human Centric AI Zinrai". We have been advancing the Research and Development on Deep Learning Techonologies with the concept of "fast", "widely used" and "easy to use". We introduce the following researchs. 1. Memory reduction method for deep neural network training2. Acceleration of a Deep Learning Framework with MPI3. Column Weight Pruning for Accelerating DNN Inferences4. Fast Algorithm Using Summed Area Tables with Unified Layer Performing Convolution and Average Pooling5. An Automated CNN Recommendation System for Image Classification Tasks1. is the research which improves the reusabilities of memories during the training phase of Deep Neural Networks. It enables us to reduce the GPU memory usage. 2. is the research which realizes the speed-up of the distributed training of Deep Neural Networks. This method utilizes MPI functions efficienlty. 3. is the research which realizes the fast recognition process of Deep Neural Networks. It reduces the amout of calculation and parameters of convolutional neural networks so that it can make use of the fast matrix calculation functionalities of the GPUs. 4. is the research about the algorithm which realizes the efficient calculation of a convolutional layer followed by an averagin-pooling layer. It performs the calculation of a pair formed by a convolutional layer and the following average-pooling layer without the computation of convolution. 5. is the research which makes it possible to suggest suitable Convolutional Neural Networks according to the given datasets. This method estimates the complexity score of the classification task as well as the classification ability score of the Convolutional Neural Networks. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Deep Learning / Neural Network / GPU / CNN / Convolutional Neural Network / Convolution / MPI / algorithm |
Paper # | CPM2017-86,ICD2017-45,IE2017-71 |
Date of Issue | 2017-10-30 (CPM, ICD, IE) |
Conference Information | |
Committee | VLD / DC / CPSY / RECONF / CPM / ICD / IE / IPSJ-SLDM / IPSJ-EMB / IPSJ-ARC |
---|---|
Conference Date | 2017/11/6(3days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Kumamoto-Kenminkouryukan Parea |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Design Gaia 2017 -New Field of VLSI Design- |
Chair | Hiroyuki Ochi(Ritsumeikan Univ.) / Michiko Inoue(NAIST) / Koji Nakano(Hiroshima Univ.) / Masato Motomura(Hokkaido Univ.) / Fumihiko Hirose(Yamagata Univ.) / Hideto Hidaka(Renesas) / Takayuki Hamamoto(Tokyo Univ. of Science) / Kiyoharu Hamaguchi(Shimane Univ.) / 渡辺 晴美(東海大) / Masahiro Goshima(NII) |
Vice Chair | Noriyuki Minegishi(Mitsubishi Electric) / Satoshi Fukumoto(Tokyo Metropolitan Univ.) / Hidetsugu Irie(Univ. of Tokyo) / Takashi Miyoshi(Fujitsu) / Yuichiro Shibata(Nagasaki Univ.) / Kentaro Sano(Tohoku Univ.) / Mayumi Takeyama(Kitami Inst. of Tech.) / Makoto Nagata(Kobe Univ.) / Kazuya Kodama(NII) / Hideaki Kimata(NTT) |
Secretary | Noriyuki Minegishi(Hiroshima City Univ.) / Satoshi Fukumoto(NTT) / Hidetsugu Irie(Kyoto Sangyo Univ.) / Takashi Miyoshi(Tokyo Inst. of Tech.) / Yuichiro Shibata(Utsunomiya Univ.) / Kentaro Sano(Hokkaido Univ.) / Mayumi Takeyama(Hiroshima City Univ.) / Makoto Nagata(e-trees.Japan) / Kazuya Kodama(Nihon Univ.) / Hideaki Kimata(Toyohashi Univ. of Tech.) / (Univ. of Tokyo) / (Panasonic) / (Nagoya Univ.) |
Assistant | / Masayuki Arai(Nihon Univ.) / Yasuaki Ito(Hiroshima Univ.) / Tomoaki Tsumura(Nagoya Inst. of Tech.) / Yuuki Kobayashi(NEC) / Hiroki Nakahara(Tokyo Inst. of Tech.) / Yuichi Akage(NTT) / Masanori Natsui(Tohoku Univ.) / Masatoshi Tsuge(Socionext) / Hiroyuki Ito(Tokyo Inst. of Tech.) / Pham Konkuha(Univ. of Electro-Comm.) / Yasutaka Matsuo(NHK) / Kazuya Hayase(NTT) |
Paper Information | |
Registration To | Technical Committee on VLSI Design Technologies / Technical Committee on Dependable Computing / Technical Committee on Computer Systems / Technical Committee on Reconfigurable Systems / Technical Committee on Component Parts and Materials / Technical Committee on Integrated Circuits and Devices / Technical Committee on Image Engineering / Special Interest Group on System and LSI Design Methodology / Special Interest Group on Embedded Systems / Special Interest Group on System Architecture |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Invited Talk] Researchs on high-speed and efficient Deep Learning technologies |
Sub Title (in English) | |
Keyword(1) | Deep Learning |
Keyword(2) | Neural Network |
Keyword(3) | GPU |
Keyword(4) | CNN |
Keyword(5) | Convolutional Neural Network |
Keyword(6) | Convolution |
Keyword(7) | MPI |
Keyword(8) | algorithm |
1st Author's Name | Takuya Fukagai |
1st Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
2nd Author's Name | Koichi Shirahata |
2nd Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
3rd Author's Name | Yasumoto Tomita |
3rd Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
4th Author's Name | Tetsutaro Hashimoto |
4th Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
5th Author's Name | Atsushi Ike |
5th Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
6th Author's Name | Masafumi Yamazaki |
6th Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
7th Author's Name | Akihiko Kasagi |
7th Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
8th Author's Name | Tsuguchika Tabaru |
8th Author's Affiliation | Fujitsu Laboratories(Fujitsu Lab. Ltd.) |
9th Author's Name | Liuan Wang |
9th Author's Affiliation | Fujitsu Research and Development Center Co., Ltd.(FRDC) |
10th Author's Name | Song Wang |
10th Author's Affiliation | Fujitsu Research and Development Center Co., Ltd.(FRDC) |
11th Author's Name | Li Sun |
11th Author's Affiliation | Fujitsu Research and Development Center Co., Ltd.(FRDC) |
12th Author's Name | Jun Sun |
12th Author's Affiliation | Fujitsu Research and Development Center Co., Ltd.(FRDC) |
Date | 2017-11-07 |
Paper # | CPM2017-86,ICD2017-45,IE2017-71 |
Volume (vol) | vol.117 |
Number (no) | CPM-275,ICD-276,IE-277 |
Page | pp.pp.39-41(CPM), pp.39-41(ICD), pp.39-41(IE), |
#Pages | 3 |
Date of Issue | 2017-10-30 (CPM, ICD, IE) |