本研究の目的は,機械が人間と自然な会話を行うために重要な能力である言語を理解する方法を実現することである.具体的には,基準となる言語資源に登録されていない単語(未定義語)の意味を提示するシステムの実現を目指す.未定義語の意味を推定するためには,ある単語から概念を想起し,さらに,その概念に関係のある様々な概念を連想できる能力が重要な役割を果たす.そこで,連想能力を表現するメカニズムである連想メカニズムを活用して,未定義語の意味を推定する方法を提案し,その有効性を評価する. / The purpose of this study is to realize a scheme to understand language, which is a significant ability for machines to have natural conversation with humans. Concretely speaking, we aim to achieve a system that can show meaning of words (undefined words) which are not registered in the reference language resource. In order to estimate the meaning of undefined words, two capability play an important role; one is recalling a concept from a word, and the other is associating various related concepts from the concept. Therefore, we propose a scheme for estimating the meaning of undefined words by using the association mechanism and evaluate its effectiveness. / 博士(工学) / Doctor of Philosophy in Engineering / 同志社大学 / Doshisha University
Identifer | oai:union.ndltd.org:doshisha.ac.jp/oai:doshisha.repo.nii.ac.jp:00027636 |
Date | 19 September 2020 |
Creators | 後藤 和人, Kazuto Goto |
Source Sets | Doshisha University |
Language | Japanese |
Detected Language | English |
Type | Thesis or Dissertation |
Format | application/pdf |
Source | https://doors.doshisha.ac.jp/opac/opac_link/bibid/BB13155935/?lang=0 |
Page generated in 0.0019 seconds