Exquisitor: Studying the Interplay Between Conversational Search and Relevance Feedback

The Exquisitor search system employs two distinct paradigms for searching media collections: conversational search and user relevance feedback. These paradigms can be used independently or cooperatively within the search process. While we posit that these techniques are complementary, as conversational search can find specific examples for relevance feedback and relevance feedback can help the user better understand the collection and lead to improved queries, empirical user behavior in real-world search scenarios remains poorly understood. To explore how these approaches interact within Exquisitor, we must gather more data about user actions during sessions, as well as their post-session feedback on the system. At IVR4B, we intend to obtain this information through user interaction logs and postsession interviews.