| Presentation | 2019-06-07 Before the Edge of Chaos Taichi Haruna, Kohei Nakajima, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We study short-term memory of discrete-time nonlinear recurrent neural networks driven by small input signals. We theoretically calculate memory capacity by combining a mean-field theory with the small input assumption. We show that the network contribution of memory capacity peaks in the parameter region which is chaotic in the absence of input signals while is ordered in the presence of input signals. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Recurrent neural networks / Reservoir computing / Memory capacity / Mean-field theory |
| Paper # | NLP2019-23,CCS2019-6 |
| Date of Issue | 2019-05-31 (NLP, CCS) |
| Conference Information | |
| Committee | CCS / NLP |
|---|---|
| Conference Date | 2019/6/7(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | machinaka campus nagaoka |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Makoto Naruse(NICT) / Hiroaki Kurokawa(Tokyo Univ. of Tech.) |
| Vice Chair | Shigeki Shiokawa(Kanagawa Inst. of Tech.) / Tetsuya Asai(Hokkaido Univ.) / Kiyohisa Natsume(Kyushu Inst. of Tech.) |
| Secretary | Shigeki Shiokawa(Hiroshima City Univ.) / Tetsuya Asai(Kanagawa Inst. of Tech.) / Kiyohisa Natsume(Nippon Inst. of Tech.) |
| Assistant | Hidehiro Nakano(Tokyo City Univ.) / Kazuki Nakada(Tsukuba Univ. of Tech.) / Hiroyasu Ando(Tsukuba Univ.) / Takashi Matsubara(Kobe Univ.) / Yutaka Shimada(Saitama Univ.) / Toshikaza Samura(Yamaguchi Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Complex Communication Sciences / Technical Committee on Nonlinear Problems |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Before the Edge of Chaos |
| Sub Title (in English) | Short-Term Memory in Echo State Networks |
| Keyword(1) | Recurrent neural networks |
| Keyword(2) | Reservoir computing |
| Keyword(3) | Memory capacity |
| Keyword(4) | Mean-field theory |
| 1st Author's Name | Taichi Haruna |
| 1st Author's Affiliation | Tokyo Woman's Christian University(TWCU) |
| 2nd Author's Name | Kohei Nakajima |
| 2nd Author's Affiliation | The University of Tokyo(Tokyo Univ.) |
| Date | 2019-06-07 |
| Paper # | NLP2019-23,CCS2019-6 |
| Volume (vol) | vol.119 |
| Number (no) | NLP-71,CCS-72 |
| Page | pp.pp.25-29(NLP), pp.25-29(CCS), |
| #Pages | 5 |
| Date of Issue | 2019-05-31 (NLP, CCS) |