Re: what research exists on visualizing search results?

How to render complex information sets in ways that make them readily accessible to people is an area of active research for a number of uses, including making digital libraries work better. The projects I am familiar with combine two fundamental approaches to the problem to get the job done. First, they work to provide a means of accurately subsetting the information space so that the user need only deal with relevant portions of it. Second, they work to represent the subset of the information space in meaningful ways.

The work I am most familiar with for accurately subsetting information spaces is that behind the IBM Digital Library product, many of the Media Miner components, and the Query by Image Content (QBIC) project. If you are interested in this part of the problem, the QBIC page has a link for contacting the researchers.

A good starting point for research in how to visualize an information set may be found on the Intelligent Graphics project at the MIT Media Lab. This page contains descriptions of various approaches researchers at the lab are taking and lists the principal researcher for each.