ChaTEL: A NOVEL VOICE COMMUNICATION SYSTEM BASED ON ANALYSING MULTIPLE TOPIC THREADS IN TEXT-BASED CHAT CONVERSATIONS

摘要

We usually have to share single-topic threads in an everyday conversation for a while and to continue talking about ideas related to the current thread. Moreover, we have to maintain a rule of turn-taking: there can only be one speaker at a given time, and the other participants should be listeners. Therefore, it is impossible for people who do not have the "turn" to talk about what they want to say, even if they think of interesting topics or good ideas. As a result, regrettably, they often forget the topics or ideas. Thus, such a daily conversation has an inefficient aspect. On the other hand, in text-based chat conversations, we can converse when we want to talk and on whatever subject we want to talk about because we do not have to follow a rule of turn-taking. In this paper we propose a novel voice communication system named "ChaTEL" that achieves multithreaded voice communication based on analysing the situations of multiple topic threads in text-based chat conversations. We confirm that ChaTEL facilitates multithreaded conversations based on user studies.