Chatbot responses can be generated using traditional rule-based conversation design or through the use of large language models (LLMs). In this paper we compare the quality of responses provided by LLM-based chatbots with those provided by traditional conversation design. The results suggest that in some cases the use of LLMs could improve the quality of chatbot responses. The paper concludes by suggesting that a combination of approaches is the best way forward and suggests some directions for future work.
|Title of host publication||Knowledge Science, Engineering and Management - 16th International Conference, KSEM 2023|
|Editors||Zhi Jin, Yuncheng Jiang, Wenjun Ma, Robert Andrei Buchmann, Ana-Maria Ghiran, Yaxin Bi|
|Place of Publication||Germany|
|Number of pages||10|
|Volume||LNCS, volume 14118|
|Publication status||Published online - 9 Aug 2023|
|Event||16th International Conference, KSEM 2023, - China, Guangzhou|
Duration: 16 Aug 2023 → 19 Oct 2023
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||16th International Conference, KSEM 2023,|
|Abbreviated title||KSEM 2023|
|Period||16/08/23 → 19/10/23|
Bibliographical noteFunding Information:
Michael McTear received support from the e-VITA project (https://www.e-vita.coach/) (accessed on 23 April 2023). Sheen Varghese Marokkie and Yaxin Bi received support from the School of Computing, Ulster University (https://www.ulster.ac.uk/faculties/computing-engineering-and-the-built-environme nt/computing).
1The e-VITA project has received funding from the European Union H2020 Pro-gramme under grant agreement no. 101016453. The Japanese consortium received funding from the Japanese Ministry of Internal Affairs and Communication (MIC), Grant no. JPJ000595.
© 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
- Conversational AI
- Conversation design
- Large Language Models