Microsoft eyes peer review for search filtering

Microsoft eyes peer review for search filtering


What your friends like will get you better search results

Microsoft is looking at exploiting information from a user's friends and business contacts to deliver better search results.

The software giant demonstrated project Nocturnal at the Microsoft Research Silicon Valley Road Show at the company's Mountain View, California office.

Nocturnal taps into the bookmarks and previously visited pages for contacts in the MSN Messenger buddy list to filter search results.

"There has been a lot of talk of social communities, but it's been hard to bootstrap the process," said Dahlia Malkhi, a researcher with Microsoft Research. "We use the buddy list to determine the social community."

Using information from users associated with the searcher will make for more accurate results, Malkhi argued.

The technology allows a search engine to push the most relevant result for an individual user to the top of the page, rather than it being buried on the second or third page.

It will also help in the battle against search engine spam where spammers try to fool engines into including irrelevant search results.

Malkhi's friends, for instance, are more likely to be looking for information about the researcher when they submit her first name, Dahlia, as a search query, whereas the general public is most likely to be looking for information on the flower.

To address privacy concerns, users will have to opt in to share their information. They can also set the breadth of their peer group, which can range from people on their buddy list or include their buddies' buddies and their buddies.

The project is currently a prototype using the MSN Live Messenger Beta. Malkhi said that the technology is ready to be launched, but could not comment on concrete plans to use it.