Generating manzai-scenario using entity mistake

Abstract

Today, many communication robots have been developed. Nevertheless, smooth communication with humans is difficult for robots. We specifically examine communication between robots based on dialogue. Here, we create a dialogue-based scenario for the robots. Our proposed scenario includes not only daily dialogue but also dialogue with humor to make the news familiar for people. Our proposed dialogue-based scenario is called a Manzai scenario. We have already proposed a means of generating Manzai scenarios automatically. Our proposed Manzai scenario consists of Introduction part, Body part, and Conclusion part. Introduction part consists of some components, one of the components is rival word extraction. There are some problem in our earlier proposed Manzai scenario creation system. Then, in this paper, we modify rival word extraction using Word2Vec. Furthermore, we modify the Introduction part based on an entity mistake. We conducted an experiment to measure the benefits of our proposed method. The experiment results underscore its benefits.