Users2Story - On the Importance of Understanding
               Searchers’ Information Needs

                                               Udo Kruschwitz
                                              University of Essex
                                               udo@essex.ac.uk




Abstract
Uncovering the storyline of events can be of interest for numerous reasons. For example, looking at it through
information retrieval glasses (with a touch of personalisation), one might want to string together the various
search requests and browsing patterns of an individual user to understand what this user might be after, what
information needs are actually hidden behind those queries with the aim of offering better information access
support. We all know what interesting stories can actually be derived like this as illustrated by the analysis of the
AOL query logs, and the “AOL search data leak” also demonstrates the problems with privacy when analysing
user’s Web search logs.
   If however we move from a Web search scenario to a professional search scenario such as search within an
organisation, then we note that privacy issues turn out to be much less of a problem. At the same time, it
becomes even more desirable to capture the search and browsing patterns of users because finding information
in this context can be very difficult and without understanding the user it can become a very frustrating exercise
for the searcher trying to find the document(s) relevant for the task at hand.
   The talk will explore this use case in some detail and in the best case will be seen as a different take on the
main theme of the workshop (and in the worst case will be judged as “off topic” with no chance of re-invitation).




Short Bio
Udo Kruschwitz is a Professor in the School of Computer Science and Electronic Engineering at the University
of Essex. His main research interest is the interface between information retrieval (IR) and natural language
processing (NLP). Professor Kruschwitz has successfully led research projects developing algorithms to turn
unstructured and partially structured textual data into structured knowledge and user/cohort models that have
been applied in a variety of applications including search, navigation and summarisation. He has furthermore
been involved as PI and Co-I in successful joint university/industry projects that turned NLP and IR research
into practical commercial applications, for example as PI on a Knowledge Transfer Partnership (KTP) project
with Signal Media. The collaboration won the Best Knowledge Transfer Partnership Project of the Year Award
at InnovateUK’s ‘Best of the Best 2015’ event, and another KTP project with Signal Media has just started.