다중 전환 대화를 위한 상황 인지 질의응답 시스템

Abstract

Recently many companies have introduced artificial intelligence chatbots. Furthermore, many recent researches have been doing to pursue natural conversation similar to human-interactive conversation and to develop sequenced conversation system which make possible not only single-turn conversation but also multi-turn conversations. However, in case of changing the sequential conversation on other subjects while having a conversation on one subject, many chatbot systems may not provide proper answers because of, lack of awareness on the context of the changed conversation, In this study, we propose a new question answering system that exploits a subject classification model, answer prediction model, and historage storage for context-aware multi-turn conversations.