Presentation | 2018-08-27 [Invited Talk] Product models and semi-supervised word segmentation Daichi Mochihashi, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | While deep learning methods have achieved revolutionary success inspeech and audio research, the impact is less significant in natural languageprocessing. This is due to the fact that the language has inherent structure, andsimple deep learning layers cannot represent such structures. Therefore, to combine speech and language it is quite important to leveragestructured statistical models for the joint models. Boltzmann machines, which is the most basic building blocks in deep learning, are product models from a statistical viewpoint. In this talk, I would like tointroduce a semi-supervised learning on product models that enablessemi-supervised word segmentation, a quite desired but not well representedtechnique for natural language processing. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | product models / semi-supervised learning / unsupervised learning / Bayesian learning / word segmentation |
Paper # | SP2018-28 |
Date of Issue | 2018-08-20 (SP) |
Conference Information | |
Committee | SP |
---|---|
Conference Date | 2018/8/27(1days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Kyoto Univ. |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | |
Chair | Yoichi Yamashita(Ritsumeikan Univ.) |
Vice Chair | Akinobu Ri(Nagoya Inst. of Tech.) |
Secretary | Akinobu Ri(Meijo Univ.) |
Assistant | Satoshi Kobashikawa(NTT) / Tomoki Koriyama(Tokyo Inst. of Tech.) |
Paper Information | |
Registration To | Technical Committee on Speech |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | [Invited Talk] Product models and semi-supervised word segmentation |
Sub Title (in English) | |
Keyword(1) | product models |
Keyword(2) | semi-supervised learning |
Keyword(3) | unsupervised learning |
Keyword(4) | Bayesian learning |
Keyword(5) | word segmentation |
1st Author's Name | Daichi Mochihashi |
1st Author's Affiliation | The Institute of Statistical Mathematics(ISM) |
Date | 2018-08-27 |
Paper # | SP2018-28 |
Volume (vol) | vol.118 |
Number (no) | SP-198 |
Page | pp.pp.29-29(SP), |
#Pages | 1 |
Date of Issue | 2018-08-20 (SP) |