The project aims to advance the state of the art in conversational search and recommendation, by developing algorithms for transparent understanding and obtaining of user interests.
The project targets three specific areas of functionality in conversational systems: (1) explicit modeling of user knowledge, (2) adaptive language usage, and (3) system initiative.
The development will take place in the application context of scholarly activities (i.e., as a conversational service related to Google Scholar and similar services), and a prototype conversational search system is to be built to serve as a platform for evaluation. However, the conclusions are intended to be general and apply to any recommendation space.
Funding: The project is partially funded by an unrestricted gift from Google
People: Krisztian Balog, Ivica Kostric