Comment 1
Comment 2
Reference:
Supporting Exploratory Information Seeking by Epistemology-based Social Search
Yuqing Mao, Haifeng Shen, Chengzheng Sun
IUI '10
Summary:
Search engines are currently limited by the use of improper keywords and evaluating the relevance of results. The authors of this paper have designed a prototype system, Baijia, that uses an epistemology-based social search on top of current exploratory information seeking (EIS). In an EIS process, a user inputs a set of queries. To form an epistemology of an EIS process, an epistemology of each query term is taken and the common results are returned. When a query is entered, the epistemology is automaticly added to the system. The user can then add selected pages to the epistemology, where other users can rank and comment on the page. As more users contribute to an epistemology, it becomes more likely that new searchers will be able to reuse it. In their preliminary studies, the authors found that Baijia invariably outperformed the AOL search engine.
Discussion:
This is certainly an interesting modification of search engines. As the rankings of results are dynamic, it is much more likely that you will get the most pertinant/up to date information. I can see some possible problems however, particularly in relation to controversial topics. It seems like users would be able to bury a relevant result that they disagreed with, reducing the likelihood that it would come up for users who would find the page informative. Still, Baijia seems like a great upgrade over current search engines.
This idea is similar to the paper I read. The name of this system is HeyStaks which is currently online. The basic idea is to allow users to create search communities and give recommendations to users when searching for specific keywords. When I read the paper, I did not think there was much research being done in this but I guess I was wrong.
ReplyDeleteThis is very interesting. I had to google that word (epistemology), but it seems like a cool idea. In my experience, crowd-sourced applications usually have a lack of information, but if enough people used this it could prove to be useful.
ReplyDeleteIt seems like that could be a very useful tool in the right hands. I'm sure existing search engines will implement this sort of thing over the next few years.
ReplyDeleteBurying results they don't like actually seems like a pretty serious concern - I wouldn't to be this to push the current model out of existence.
ReplyDeleteBurying results does seem like a difficult issue to approach. It's a cool idea and who knows, we might see it augment search engines one day.
ReplyDeleteI share your concerns about people being able to bury a particular result based on their disagreement. It would be interesting to see if a mechanism to avoid this could be designed.
ReplyDelete