| Presentation | 2017-10-12 Entanglement Entropic Convolutional Neural Network Shu Eguchi, Masaru Tanaka, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The neural network used for machine learning is an extract that extracts information necessary for classification from enormous data so as not to undergo overlearning. However, when extracting information necessary for classification, it also extracts unnecessary information for classification. In this article, we report on the results of experiments conducted to determine whether an algorithm to maintain important information is possible while reducing unnecessary information for classification. The model discussed in this paper is a convolution neural network based on an entanglement entropy (EECNN: Entanglement Entropic Convolutional Neural Network), which uses a physical approach to the machine learning field. The calculation of the entangement entropy using the probability amplitude obtained from input to the layer or output from the layer. The entangement entropy is an example where the original image can be sufficiently approximated by reconstructing using several singular values after the singular value decomposition, if the probability amplitude is not random noise. Maintaining important information while reducing information by restoring the original probability amplitude from singular values based on the results of entropy. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | quantum mechanics / machine learning / entanglement entropy / singular value decomposition / convolutional neural network |
| Paper # | PRMU2017-73 |
| Date of Issue | 2017-10-05 (PRMU) |
| Conference Information | |
| Committee | PRMU |
|---|---|
| Conference Date | 2017/10/12(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Shinichi Sato(NII) |
| Vice Chair | Hironobu Fujiyoshi(Chubu Univ.) / Yoshihisa Ijiri(Omron) |
| Secretary | Hironobu Fujiyoshi(AIST) / Yoshihisa Ijiri(NAIST) |
| Assistant | Masato Ishii(NEC) / Yusuke Sugano(Osaka Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Pattern Recognition and Media Understanding |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Entanglement Entropic Convolutional Neural Network |
| Sub Title (in English) | |
| Keyword(1) | quantum mechanics |
| Keyword(2) | machine learning |
| Keyword(3) | entanglement entropy |
| Keyword(4) | singular value decomposition |
| Keyword(5) | convolutional neural network |
| 1st Author's Name | Shu Eguchi |
| 1st Author's Affiliation | Fukuoka University(Fukuoka Univ.) |
| 2nd Author's Name | Masaru Tanaka |
| 2nd Author's Affiliation | Fukuoka University(Fukuoka Univ.) |
| Date | 2017-10-12 |
| Paper # | PRMU2017-73 |
| Volume (vol) | vol.117 |
| Number (no) | PRMU-238 |
| Page | pp.pp.61-66(PRMU), |
| #Pages | 6 |
| Date of Issue | 2017-10-05 (PRMU) |