III: Small: Search Assistance Using Search Trails

点击翻译按钮获取中文摘要

Current search systems are effective in helping users complete simple search tasks, but provide less support for complex tasks. When users search for information, they create "search trails" based on their interactions with the search system. These trails have valuable information that could benefit a future searcher working on a similar task and may include the queries issued, results clicked, pages viewed, pages bookmarked, and annotations made by a previous searcher. Currently, many search engines use these activity traces to improve their search algorithms and results. However, current approaches are limited; they may lead to indirect benefits, but do not consider the full potential of search trails as a direct form of search assistance. This research will develop and evaluate systems that automatically display relevant search trails as a form of search assistance to users. Using search trails for search assistance has the potential to improve a broad range of systems, including web search engines used by millions, digital libraries, and enterprise and website-specific search engines that serve users with similar goals and needs. The outcomes of this project will expand the accessibility of search across a wide range of domains. The research project will produce data that will allow other researchers to develop and evaluate their own solutions and tools that will facilitate others to perform large-scale studies of search behavior. Insights gained from the studies will also be of interest to researchers in other fields such as psychology and education. Prior research on search trails has suggested the usefulness of search trails, but has not answered key research challenges required to design and implement them. The system needs to predict when to display search trails to a user, which trails to display, and how to display them in a way that supports the user's goal. These challenges will be addressed in three phases. Phase 1 will determine which factors of the user, task, and system influence whether a searcher wants help, for what purpose, and whether they are able to gain useful information. These results will have implications for the design of assistance systems across a variety of domains. Phase 2 will develop models for predicting when to show trails to a user based on user and task features, as well as behavioral measures that indicate whether a searcher is having difficulty. The insights gained will help the development of search engines that are more user and task-aware. Finally, Phase 3 will develop models for predicting which trails to show for the current search session. Being able to match search sessions based on the user's higher-level goal has direct implications to other information retrieval tasks such as document ranking, query suggestion, and aggregated search. We will use learning-to-rank algorithms to combine features that measure the similarity between the current search session and the candidate trail, and the information content in the trail.