Presentation 1996/7/18
An Automatic Learning Method of Grammar Rules for Spontaneous Speech by Adding Bigram
Koji OTANI, Seiichi NAKAGAWA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we describe an automatic learning method of the grammar rules for improving coverage of the sentences. It is possible to analyze the sentences that can not be accepted by preregistered production rules. When a sentence which is not accepted by production rules is inputted to the system, the system learns rules using this input sentence. As a result, the coverage of the sentence can be improved. Except for regisering unknown words, there are three methods for learning the grammar. One is to add new production rules to the original CFG. The second is to add word pairs. The last is to add nonterminal symbol pairs. In the registering unknon words or not, we evaluate the coverage of sentences and perplexity using above methods.And we evaluate perplexity using bigram instead of word pairs.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) learning of grammar / CFG / word pair / nonterminal symbol pair / bigram / coverage, perplexity
Paper # NLC96-12
Date of Issue

Conference Information
Committee NLC
Conference Date 1996/7/18(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 Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An Automatic Learning Method of Grammar Rules for Spontaneous Speech by Adding Bigram
Sub Title (in English)
Keyword(1) learning of grammar
Keyword(2) CFG
Keyword(3) word pair
Keyword(4) nonterminal symbol pair
Keyword(5) bigram
Keyword(6) coverage, perplexity
1st Author's Name Koji OTANI
1st Author's Affiliation Department of Information and Computer Sciences Toyohashi University of Technology()
2nd Author's Name Seiichi NAKAGAWA
2nd Author's Affiliation Department of Information and Computer Sciences Toyohashi University of Technology
Date 1996/7/18
Paper # NLC96-12
Volume (vol) vol.96
Number (no) 157
Page pp.pp.-
#Pages 8
Date of Issue