A Framework for Task Retrieval in Task-Oriented Service Navigation System

Abstract

Mobile access to the Internet is increasing drastically, and this is raising the importance of information retrieval via mobile units. We have developed a task-oriented service navigation system [6] that allows a user to find the mobile contents desired from the viewpoint of the task that the user wants to do. However, the user is still faced with the problem of having to select the most appropriate task from among the vast number of task candidates; this is difficult due to the fact that mobile devices have several limitations such as small displays and poor input methods. This paper tackles this issue by proposing a framework for retrieving only those tasks that suit the abstraction level of the user’s intention. If the user has settled on a specific object, the abstraction level is concrete, and tasks related to the handling of the specific object are selected; if not, tasks related to general objects are selected. Finally, we introduce two task retrieval applications that realize the proposed framework. By using this framework, we can reduce the number of retrieved tasks irrelevant to the user; simulations show that roughly 30% fewer tasks are displayed to the user as retrieval results.