Presentation 2006-05-26
A model of word sequence prediction by using bigram
Yoshihisa SHINOZAWA, Kengo UEHARA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Elman proposed simple recurrent network which is a model of language acquisition. Elman showed that SRN learns to predict the next word of the sentences and can acquire grammatical concepts and meanings. We think that it is difficult for SRN to learn the sentences which contain a number of words, especially to learn new words. We improve SRN and propose a model of word sequence prediction which learns new words additionally. We propose how to learn to predict the next words by distributed networks, whose structure is decided by using bigram. We evaluate our model with learning of next word prediction.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Simple Recurrent Network / Word acquisition / Word sequence prediction / Bigram
Paper # NC2006-8
Date of Issue

Conference Information
Committee NC
Conference Date 2006/5/19(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A model of word sequence prediction by using bigram
Sub Title (in English)
Keyword(1) Simple Recurrent Network
Keyword(2) Word acquisition
Keyword(3) Word sequence prediction
Keyword(4) Bigram
1st Author's Name Yoshihisa SHINOZAWA
1st Author's Affiliation Department of Administration Engineering, Faculty of Science and Technology, Keio University()
2nd Author's Name Kengo UEHARA
2nd Author's Affiliation Laboratory of Administration Engineering
Date 2006-05-26
Paper # NC2006-8
Volume (vol) vol.106
Number (no) 79
Page pp.pp.-
#Pages 6
Date of Issue