発表番号P-3
発表タイトル 大規模言語モデルとロボットを用いたアクティブラーニングにおけるリヴォイシングタイミングの研究
筆者氏名・所属 大場 璃々 (慶應義塾大学理工学部)
松岡 竜輝 (慶應義塾大学大学院理工学研究科)
井上 南 (慶應義塾大学大学院理工学研究科)
今井 倫太 (慶應義塾大学理工学部)
アブストラクト In the field of education, active learning with robots is becoming increasingly popular as a means to enhance children's proactivity. However, there is a problem that children might not develop their critical thinking skills if they directly ask robots for answers. In this paper, we propose the use of a robot named RevoicingRobot, which vocalizes revoicing sentences generated by large-scale language models during group learning activities. “Revoicing” means listeners respond to speakers by paraphrasing or building upon their statements.The robot determines response tags from the conversation and performs revoicing accordingly. We test whether discussions become more lively when students solve problems with the robot.
論文 PDFファイル