Jie Zou will defend his PhD thesis entitled: Improving Search and Recommendation by Asking Clarifying Questions. Supervisor: Prof. Dr E. Kanoulas, Co-Supervisor: Prof. Dr M. de Rijke.
The work in this thesis provides a series of models and online user studies for search and recommender systems that construct and use CQs. In the first part of the thesis, we focus on CQ-based search and recommendation algorithms. We first propose a novel CQ-based document search algorithm, Sequential Bayesian Search based method for Technology-Assisted Review (SBSTAR) and its extension, to efficiently retrieve the last few, but significant, relevant documents. The algorithm sequentially selects and directly asks CQs to users about the presence or absence of an entity in the missing relevant documents. Next, we propose a novel CQ-based product search algorithm, Question-based Sequential Bayesian Product Search (QSBPS), to effectively locate products that users are looking for. The model is based on a duet learning framework that learns product relevance and the reward of the potential CQ to be asked to the user. Furthermore, we propose a novel CQ-based recommendation algorithm, Question-based recommendation (Qrec), to assist users to find items interactively. The model is first trained offline by a novel matrix factorization algorithm, and then iteratively updates the user and item latent factors online based on the user's answers. We experimentally test the performance of the proposed models and demonstrate that they outperform state-of-the-art baselines.
In the second part of the thesis, we focus on the evaluation of CQ-based search and recommendation. We first conduct an online user study by deploying a CQ-based system in the domain of online retail, to understand to what extent users can answer CQs. We explore the user willingness, ability, and user perception to provide answers to CQs. We find that users are willing to answer a good number of the system generated questions and most users answer questions until they reach the target product. Users also provide incorrect answers while answering CQs, and most users are positive towards CQ-based systems. Next, we conduct a large user study on web search to understand the impact of CQ interactions on user search behavior and satisfaction, and how users interact with different levels of quality of CQs under different user perceptions and conditions. We find that user interactions with high quality CQs improve user performance and satisfaction, while low and mid quality CQs are harmful. We also observe that user engagement, and therefore the user need for CQ support, is affected by various factors, such as search result quality or perceived task difficulty.