Presentation | 2020-03-02 Dimension reduction without multiplication in machine learning Nobutaka Ono, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | In this study, we propose a dimension reduction method for machine learning by only selecting elements without multiplication. In machine learning, the dimension reduction is an important preprocessing that reduces unnecessary model parameters, and it alleviates over-fitting and improves learning speed. One of the basic and frequently used methods is principal component analysis (PCA). However, since PCA requires matrix multiplication, it is not suitable for a system with limited computational power such as hearing aids and embedded devices. The number of times of multiplication in PCA itself may cause a large calculation load. In this study, to reduce the amount of calculation in dimension reduction, we consider dimension reduction only by element selection. What is essential at this time is which elements of the input vector are selected. In this study, we consider an objective function defined as the reconstruction loss of a linear autoencoder, and this is formulated as a discrete optimization problem that selects the element that minimizes it. Also, we propose a method to solve this problem by sequentially replacing elements chosen so that the objective function becomes smaller. The calculation of the objective function includes the inverse matrix operation. An algorithm that significantly reduces the amount of computation using the inverse matrix lemma is described. Preliminary experimental results for video show the effectiveness of this method. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | dimension reduction / machine learning / element seelction / multiplication / computational cost |
Paper # | EA2019-104,SIP2019-106,SP2019-53 |
Date of Issue | 2020-02-24 (EA, SIP, SP) |
Conference Information | |
Committee | SP / EA / SIP |
---|---|
Conference Date | 2020/3/2(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Okinawa Industry Support Center |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Hisashi Kawai(NICT) / Kenichi Furuya(Oita Univ.) / Naoyuki Aikawa(TUS) |
Vice Chair | Akinobu Ri(Nagoya Inst. of Tech.) / Suehiro Shimauchi(Kanazawa Inst. of Tech.) / Shigeto Takeoka(Shizuoka Inst. of Science and Tech.) / Kazunori Hayashi(Osaka City Univ) / Yukihiro Bandou(NTT) |
Secretary | Akinobu Ri(Kyoto Univ.) / Suehiro Shimauchi(Waseda Univ.) / Shigeto Takeoka(NHK) / Kazunori Hayashi(Univ. of Tokyo) / Yukihiro Bandou(Hiroshima Univ.) |
Assistant | Tomoki Koriyama(Univ. of Tokyo) / Yusuke Ijima(NTT) / Keisuke Imoto(Ritsumeikan Univ.) / Daisuke Morikawa(Toyama Pref Univ.) / Kenjiro Sugimoto(Waseda Univ.) |
Paper Information | |
Registration To | Technical Committee on Speech / Technical Committee on Engineering Acoustics / Technical Committee on Signal Processing |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Dimension reduction without multiplication in machine learning |
Sub Title (in English) | |
Keyword(1) | dimension reduction |
Keyword(2) | machine learning |
Keyword(3) | element seelction |
Keyword(4) | multiplication |
Keyword(5) | computational cost |
1st Author's Name | Nobutaka Ono |
1st Author's Affiliation | Tokyo Metropolitan University(TMU) |
Date | 2020-03-02 |
Paper # | EA2019-104,SIP2019-106,SP2019-53 |
Volume (vol) | vol.119 |
Number (no) | EA-439,SIP-440,SP-441 |
Page | pp.pp.21-26(EA), pp.21-26(SIP), pp.21-26(SP), |
#Pages | 6 |
Date of Issue | 2020-02-24 (EA, SIP, SP) |