The EntityCube may sound like the title of a new Keanu Reeves movie, but actually it’s a new research prototype from Microsoft for exploring object-level search technologies that automatically summarizes the web for entities such as people and organizations that have some web presence. It has so far extracted information from 3 billion web pages. Born in the Microsoft Research Asia group as Renlifang, the language tool is said to have proven popular there, getting millions of daily page views during peak days.

As Microsoft describes it, “EntityCube generates summaries of Web entities from billions of public Web pages that contain information about people, locations, and organizations, and allows for exploration of their relationships. For example, users can use EntityCube to find an automatically generated biography page and social-network graph for a person, and use it to discover a relationship path between two people.”

Semanticweb.com took EntityCube for a spin, starting – fittingly enough -- with a search on Bill Gates. A Name Disambiguation feature lets you choose that you are interested in the Bill Gates who is the Chairman of Microsoft, not the American frontiersman and fortune hunter of the Klondike Gold Rush or a gentleman who provides LAN support at a technology vendor. Explore his social network graph far enough and you can pull up some pretty interesting connections – for instance, news reports have him noted as a guest at Tiger Woods 2004 wedding. Most of these insights are drawn from articles that have somehow made a link between Gates and other figures (including dead ones like Isaac Newton), locations or organizations; it’s not actually mining his Facebook connections, for instance, which are private, or one of his Twitter accounts for followers. As a prototype Microsoft acknowledges that some names and relationships may be incorrect, and if you do check out Mr. Gates’ SNS accounts you’ll find that the name disambiguation hasn’t quite carried over – all the Twitter and Facebook spoof accounts are there. Whether any of them is real is doubtful as last time I checked Microsoft wasn’t commenting about whether the former CEO had any real plans to start Twittering.

But enough about Bill Gates. Where EntityCube can be more fun is when you’re trying to dig deeper into the lives of average and everyday folks, like us. I loved checking out its copyrighted Guanxi map that comes up when I did a search on my own name, for instance. It lets you see all your connections in a visual way. You can click on anyone in the list to discover their own connections and details about them that can take you down new paths to explore. With the Six Degrees of Separation feature, you can also put in two people’s names to discover a relationship path – and yes, a line runs from me to Craig Conway (former CEO of PeopleSoft) to Larry Ellison to Bill Gates, based on articles I’ve done in the past.

I was a bit disappointed to see some pieces of my stats still missing, but I’ll attribute that to EntityCube’s still having a few more billion web pages to explore. Also amusing and telltale of the work yet to be done to perfect entity extraction: EntityCube has me connected to Alfred P. West Jr., who is indeed a real person (chairman of the board and CEO of SEI Investments Co.). But my actual connection, if you want to call it that, is to a building named after the man rather than the man itself – the Alfred P. West Jr. Learning Labs which featured in a story I wrote about Wharton.

Users can also try out the system to explore topics based on keywords, or click into categories such as sports and entertainment. Type in “jobs report”, for instance, and you’ll get Bing results of news on the latest stats on the employment situation, job creation investments, etc.. as well as a list of people, authors, conferences and so on affiliated with the situation (alas, clicking on these individuals or objects doesn’t always make that relationship immediately clear).

There’s also an Academic version of EntityCube with a list of top-ranked papers by domain, as well as authors, conferences and journals. From there you can also search on specific individuals or institutions to find the news, individuals, and other objects related to them.

The Guanxi map is very interesting when viewed in light of institutions, academic or otherwise – and yes, it does plot a course from General Motors to Michael Moore (and Tiger Woods, too).

Exploring the Bill Gates connection.