Evaluation of learning efficiency for knowledge acquisition of linguistic symbol through dialogues using image and voice media

摘要

There are many studies about acquiring linguistic symbol on computer. But much of them take meanings or relations of linguistic symbol as based knowledge, and learn knowledge by symbolic process. Then, the purpose of this study is knowledge acquisition of linguistic symbol. This study has proposed the image and voice which is unsymbolic information acquiring linguistic symbol model by multi-agent using dialog. But the only way to judge whether computer learns what human wants was to suppose the image and voice from computer. Thus, by observing the phenomena on the acquiring process directly, we evaluate the linguistic symbol quantitatively and show a acquiring concept with dialog is more effective than with general instructing.