Humor can create a pleasant atmosphere, ease tension, and enhance the overall interaction experience. Yet effective humor is difficult to perform: it is spontaneous, context-sensitive, and not everyone is naturally skilled at using it. Avatar robots, however, offer new possibilities. By allowing operators to transcend spatial distance and, importantly, to compensate for limitations in human expressive ability, they provide a promising platform for supporting humorous interaction. In this work, we design a support system that enables operators to realize humor-based interaction through an avatar robot. First, we propose a general framework for humor-enabled interaction that can be applied across diverse real-world scenarios. Second, we explore the capacity of large language models (LLMs) to interpret subtle humorous context and develop a humor support system that enhances the operator’s expressive capability. Finally, we deploy the system on an avatar robot and conduct a field study in a shopping mall to evaluate its effectiveness in dynamic human–robot interactions.
| 氏名 | コース | 研究室 | 役職/学年 |
|---|---|---|---|
| Zhao Yuxin | 社会情報学コース | 神田研究室 | 博士3回生 |