Novel Query Suggestions

Query auto-completion (QAC) is one of the most recognizable and widely used services of modern search engines. Its goal is to assist a user in the process of query formulation. Current QAC systems are mainly reactive. They respond to the present request using past knowledge. Specifically, they mostly rely on query logs analysis [11, 10, 12] or corpus terms co-occurrences [8] and rank suggestions according to their similarity with the partial user query, their past popularity, or their temporal dynamics features (e.g. trends, bursts, seasonality in query popularity) [9]. Consequently, a suggestion to be recommended by the QAC system must be preceded with a substantial users’ interest and ipso facto must be an old information. However, a growing amount of people turns to search engines to find novel information, that is emergent or recently created (not redundant) one. Conventional QAC systems are thus unable to fulfill the increasingly real-time needs of the users. In this work-in-progress report, we introduce a new approach to QAC — the system filtering out potentially novel information and proactively delivering it to the users. It aims at providing the users with some novel insight. Thus, it caters for their open-ended or persistent and increasingly real-time information needs. The preliminary method proposed in this paper to evaluate this approach forms time specific suggestions based on a comparison of two corpora constantly being updated with new data from chosen sources. An unsupervised and language-independent algorithm rely∗Ilona Nawrot is also associated with Poznań University of Economics, al. Niepodleg lości 10, 61-875 Poznań, Poland. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than the author(s) must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]. Web-KR’14, November 3, 2014, Shanghai, China. Copyright is held by the owner/author(s). Publication rights licensed to ACM. ACM 978-1-4503-1606-4/14/11 ...$15.00. http://dx.doi.org/10.1145/2663792.2663799 . ing on relative novelty of terms co-occurrences is used to generate suggestions. The initial experimental results demonstrate the effectiveness of the approach in recommending queries leading to novel information. Therefore, they prove that such a system can enhance the exploratory power of a search engine and support the proactive information search.