Reproduce “Using LLMs to Investigate Correlations of Conversational Follow-up Queries with User Satisfaction” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hyunwoo Kim, Yoonseo Choi, Taehyun Yang, Honggu Lee, Chaneon Park, Yongju Lee, Jin Young Kim, Juho Kim
Researchers from KAIST, Seoul National University, and Naver Corporation investigated how follow-up queries in conversational search correlate with user satisfaction. They developed a two-dimensional taxonomy to classify these queries and used an LLM-powered system to analyze patterns, revealing that specific query types predict user satisfaction and highlighting common user behaviors like domain exploration and requesting additional information.
Key insights
- A comprehensive taxonomy reveals 18 unique follow-up query patterns in conversational search, indicating more complex user interactions than traditional search.
- User intent to 'Explore Domain' and the action of 'Requesting Additional Information' are the most frequent follow-up patterns, highlighting the exploratory nature of conversational search.
- Specific follow-up query patterns directly correlate with user satisfaction levels, offering actionable insights for system improvement and evaluation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.