Lightning Talk - Looking for the Movie Seven or Sven from the Movie Frozen?

As one of the largest communities that search for online resources, children are introduced to the Web at increasingly young ages [1]. However, popular search tools are not explicitly designed with children in mind nor do their retrieved results explicitly target children. Consequently, many young users struggle in completing successful searches, especially since most search engines (SE) do not directly support, or offer weak support, for children’s inquiry approaches [2]. Even though children, as inexperienced users, struggle with describing their information needs in a concise query, they still expect SEs to retrieve relevant information in response to their requirements. As part of their capabilities, SEs often suggest queries to aid users in better defining their information needs. In fact, a recent study conducted by Gossen et al. [3] shows that children pay more attention to suggested queries than adults. Unfortunately, these suggestions are not specifically tailored towards children and thus need improvement [5]. While there exist multiple query suggestion modules, only few specifically target children. To address this problem, along with a need for more children-related tools, we rely on ReQuIK (Recommendations based on Query Intention for Kids), a query suggestion module tailored towards 6-to-13 year old children (introduced in [4]). ReQuIK informs its suggestion process by applying (i) a strategy based on search intent to capture the purpose of a query [1], (ii) a ranking strategy based on a wide and deep neural network that considers both raw text and traits commonly associated with kid-related queries, (iii) a filtering strategy based on the readability levels of documents potentially retrieved by a query to favor suggestions that trigger the retrieval of documents matching children’s reading skills, and (iv) a content-similarity strategy to ensure diversity among suggestions.

For assessing the quality of the system, we conducted initial offline and online experiments based on 591 queries written by 97 children, ages 6 to 13. The results of this assessment verified the correctness of ReQuIK’s recommendation strategy, the fact that it provides suggestions that appeal to children and ReQuIK’s ability to recommend queries that lead to the retrieval of materials with readability levels that correlate with children’s reading skills

To the best of our knowledge, ReQuIK is the only available system that can be coupled with SEs to generate query recommendations for children, favoring those that lead to easier-to-read, child-related resources, which can improve SEs’ performance. The design of the proposed tool explicitly considers different patterns children use while searching the Web to adequately capture the intended meaning of their original queries. For example, if a child submits the query “elsa”, ReQuIK aims to prioritize query suggestions such as “elsa coloring papers” or “elsa dress up games” that correlate better with topics of interest to children rather than “elsa pataky”, as suggested by Google, which is more appealing to mature users. Other contributions of our work include a novel ranking model inspired by a deep-and-wide architecture that, while successfully applied for ranking purposes [6], has never been used in the query suggestions domain; a strategy to overcome the lack of queries written by children by taking advantage of general purpose children-oriented phrases; and a newly-created dataset [4].