DFKI-LT - Dissertation Series

Vol. XXXVIII

Sven Schmeier: Exploratory Search on Mobile Devices

ISBN: 978-3-86223-136-2
166 pages
price: € 12

order form

The research field Exploratory search, embedded in the field of Human Computer Interaction (HCI), aims for a next generation of search interfaces beyond the document centered Google-like approaches. New interfaces should support users to find information even if their goal is vague, to learn from the information, and to investigate solutions for complex information problems.

The goal of this thesis is to provide a general framework (MobEx ) for exploratory search especially on mobile devices.

The central part is the design, implementation, and evaluation of several core modules for on-demand unsupervised information extraction (IE) well suited for exploratory search on mobile devices and creating the MobEx framework. These core processing elements, combined with a multitouchable user interface specially designed for two families of mobile devices, i.e. smartphones and tablets, have been finally implemented in a research prototype. The initial information request, in form of a query topic description, is issued online by a user to the system. The system then retrieves web snippets by using standard search engines. These snippets are passed through a chain of the already mentioned NLP components which perform an on-demand or ad-hoc interactive Query Disambiguation, Named Entity Recognition, and Relation Extraction task. By on-demand or ad-hoc we mean the components are capable to perform their operations on an unrestricted open domain within special time constraints. The result of the whole process is a topic graph containing the detected associated topics as nodes and the extracted relationships as labelled edges between the nodes. The Topic Graph is presented to the user in different ways depending on the size of the device she is using. It can then be further analyzed by users so that they can request additional background information with the help of interesting nodes and pairs of nodes in the topic graph, e.g., explicit relationships extracted from Wikipedia or extracted from the snippets as well as conceptual information of the topic in form of semantically oriented clusters of descriptive phrases. Various evaluations have been conducted that help us to understand the chances and limitations of the framework and the prototype.