It is not just the product teams at Microsoft that are focusing heavily on search. The Microsoft Research unit is doing so, as well.

To wit, Microsoft Research (MSR) recently reorganized its Redmond labs by creating four new research teams. A text mining, search and navigation team is prominent among the newly minted groups.

The other new MSR teams include a communications/collaboration group; knowledge tools group; and network embedded computing group.

At the same time, MSR Asia recently announced that it also is reorganizing and reprioritizing. MSR Asia created a handful of new research teams. Among them: A Web search and mining team.

Microsoft announced the formation of the four new Redmond-based teams at the end of August, roughly around the time Microsoft promoted Henrique "Rico" Malvar to director of MSR Redmond. Previously, Malvar was a senior researcher and manager with MSR's Signal Processing Group.

MSR's text mining/search group works closely with the Windows, SQL Server, MSN and other Microsoft product teams that are focused on beating Google and other search competitors. The group is headed by Eric Brill.

"We've been doing research in this space in MSR since Day 1," especially in MSR's Cambridge (UK) and Beijing labs, said Kevin Schofield, general manager of strategy and communications for MSR.

The Redmond text mining/search team will be "a new hub for search," Schofield said. "And our investment in this area will only grow. That's where the industry is headed."

The text mining, search and navigation group is focusing on helping users "find information more effectively in large text collections, such as the Web, help and support sites, discussion groups, and intranets," according to the mission statement on the new text mining/search site.>

"A search engine should be more helpful than merely delivering a large list of documents to a user. We are building prototypes that give users a better search experience by allowing users to more effectively navigate results and have a dialog with the search engine to find what they're looking for.

"Often a user is searching for information, rather than documents. We are working on automatic question answering as a means to directly deliver the requested information to a user, rather than returning a huge set of documents that might contain the requested information," the text on the site continued.
MSR has a strong focus on natural-language processing and synthesis in the text/search group, Schofield said.

The MSR Asia Web search and mining team has a slightly different focus.

"Our goal is to bring the current Web search to the next level by applying data mining, machine learning, and knowledge discovery techniques to information analysis, organization, retrieval, and visualization," according to the search team's Web site.

MSR Asia is working on techniques to "datamine deep Web structure to enable one-stop search of multiple online databases, as well as extract page layout structure to improve link analysis and relevance ranking," the site continued.

" In contrast with the current Web search methods that essentially do document-level ranking and retrieval, we are exploring a new paradigm to enable Web search at the object level," the site said.

The MSR team also is investigating the area of mobile search.

(This is an edited version of an article which appeared in the September 2, 2004, issue of the Microsoft Watch newsletter. Want to see what other Microsoft news nuggets you might have missed? Sign up today for a free two-week trial subscription to Microsoft Watch.)