@techreport{oai:ipsj.ixsq.nii.ac.jp:00234302, author = {松尾, 三央子 and 原田, 拓実 and 松尾, 一志 and 荒井, 研一 and 小林, 透 and 中村, 典生 and Mioko, Matsuo and Takumi, Harada and Kazushi, Matsuo and Kenichi, Arai and Toru, Kobayashi and Norio, Nakamura}, issue = {1}, month = {May}, note = {日本で生まれ育った一般的な日本人の児童は,小学生から英語を習い始める.小学校における英語教育では,特にコミュニケーション能力を育てることが重要視されている.そのため,著者らは,文部科学省の学習指導要領に沿って英会話学習ができる小学生向けの英会話学習システムの開発を行っている.これまでに開発してきた英会話学習システムでは,会話の制御にルールベースのチャットボット Watson Assistant のみを用いていたため,ユニットボタンを押すたびに,どの児童に対しても同じ質問が繰り返され,児童ごとにパーソラナイズされた質問ができない課題があった.そこで,本研究では,児童の答えた会話履歴に応じて質問を動的に生成する応答的エージェントを適用させた会話制御手法を提案する.「応答的」とは,子どもの応答を重視して実践する保育のことで,こどもの成長・発達に対する有効性が認められている.具体的には,会話ベース AI(Watson Assistant)と連想ワード AI(グラフデータベース Neo4j)の 2 つの AI と児童毎の会話の履歴を制御することで,児童の会話履歴を理解し,学習指導要領準じて質問パターンを変化させる児童に寄り添う英会話学習システムを実装した.本論文では,実際に児童を対象に評価を行い,その有用性を検証した., Japanese children generally begin learning English in elementary school. English education in elementary school places particular importance on developing communicative competence. This research is developing an English conversation learning system for elementary school students. For this reason, the authors are developing an English conversation learning system for elementary school students that enables them to learn English conversation in accordance with the curriculum guidelines of the Ministry of Education, Culture, Sports, Science and Technology. The English conversation learning system we have developed so far uses only the rule-based chatbot Watson Assistant to control the conversation, so the same questions are repeated for each child every time the unit button is pressed. In this study, we propose a conversation control method that applies a responsive agent that dynamically generates questions based on the conversation history of children's answers. The term "responsive" refers to childcare practices that emphasize children's responses and are recognized as effective for children's growth and development. Specifically, by controlling the conversation history of each child with two AIs, a conversation-based AI (Watson Assistant) and an associative word AI (Neo4j, a graph database), we implemented an English conversation learning system that understands the child's conversation history and changes question patterns according to the curriculum guidelines. We implemented this system. In this paper, we actually conducted the evaluation on children and verified its usefulness.}, title = {小学生英会話学習システムにおけるパーソラナイズ化の検討}, year = {2024} }