발간년도 : [2024]
| 논문정보 |
|
| 논문명(한글) |
[Vol.19, No.1] How Long-Term Memory in Human-Agent Interactions Affects User Experience |
|
| 논문투고자 |
Eun-Sol Kim, Young-Jae Yoo, Ji-Soo Han, Chang-Hoon Oh |
|
| 논문내용 |
Recent advances in conversational agents have incorporated "long-term memory" technology to remember conversations with users. Whether or not an agent has long-term memory is expected to be an important factor in the user experience. The effect of memory may be different depending on the type of conversation of the agent. Therefore, this study aims to explore the user experience of whether an agent remembers or not depending on the type of dialog. The experiment was a 2x2 mixed design with agent's remembering and dialog type as independent variables to measure likeability, closeness, psychological resistance, and trust. The chatbot was simulated using the Wizard of OZ technique and was divided into a primary and secondary dialog. The results showed that there was a significant difference in intimacy based on the agent's dialog type, and an interaction effect between the agent's dialog type and memory was found in trust. The significance of this study is that it investigated the user experience of whether an agent remembers or not depending on the type of conversation, at a time when the number of interactive agents incorporating long-term memory technology is increasing. In addition, we expect that the results of this study will provide practical help in planning and developing real-world conversational agent systems. |
|
| 첨부논문 |
|
|
|
|
|
자료실 > 논문지 
