Paper Reading 21 – Supporting Exploratory Information Seeking by Epistemology-based Social Search

An example entry in the system.

Comments
Paola Garza
Patrick Frith

Reference Information
Supporting Exploratory Information Seeking by Epistemology-based Social Search
 - Yuqing Mao, Haifeng Shen, and Chengzheng Sun
 - IUI 2010, Hong Kong, China

Summary
This paper deals with solving the issues of effective exploratory information seeking (EIS).  EIS refers to searches where users are searching for specific information on a subject or topic.  Initially, these users tend to not have all the information they need to yield the best search results.  Suggesting keywords like Google uses is one method of improving this problem.  But, they proposed to use a social search approach to help expand on past searches.

With their system called “Baijia,” users can perform a search and then attach applicable sites to the search query they found in their search.  This helps future users by being able to not only see a possible search query similar to theirs, but they can also see what kind of results this query might yield.  The key difference they point out between their approach and similar ones is they allow other to refine these EIS social searches.  If a certain query is close or does not quite yield the results someone wants, this other person can also expand and add to the stored query.  Allowing these feature further expand on the effective of the original query, and this will also help future people improve their searching even more than before.

This is how the entry appears from new searches.

Compared to other options available, the authors found their implementation to be far superior to conventional search engines in supporting EIS needs.  They were able to find the desired information quicker and more efficiently.  The added ability to alter the social search entries also helped improve their solution to EIS.  They plan to conduct user studies to further evaluate their approach and collect information about the application’s interface.

Discussion
I tend to have fairly successful searches when I am looking for information using Google.  However, I can certainly see how this could help a great deal in improving time spend searching for certain types of information.  For items that are extremely difficult to find, it would help a lot to have groundwork laid out by people who might have conducted the same search in the past.  I would definitely like to see the results of their testing and see how practical this approach is over a broad selection of searches.  If it’s still very successful, I’d love to try their application out at some point.