Users tend to click on results ranked highly by search engines much more often than those ranked lower. How do you train a search engine using click data / search logs without this bias? I.e. you don't want to teach the search engine that the results that are currently ranked highly should necessarily continue to be ranked highly just because they were frequently clicked.
$\begingroup$
$\endgroup$
$\begingroup$
$\endgroup$
Have you seen this paper? Optimizing Search Engines using Clickthrough Data
I stumbled upon this the other day, and I'm still reading through it, but the author attempts to deal with the problem you describe. You may also find Improving Web Search Ranking by Incorporating User Behavior Information useful.
