Semantic Reality (Microsoft Acquires Powerset)

Fact: At last, the public announcement this afternoon of one of the most-rumored secrets in tech: Microsoft is acquiring Powerset, taking us one major step forward in semantic technologies. 

Analysis: There’ll be plenty of analysts looking at this, and I expect the acquisition will get a lot of buzz just as Powerset did originally when launched.  After all, Microsoft is buying a company which was called a “Google-Killer” by everyone from the New York Times to various esoteric search-technology blogs.  [Update: it’s already started on TechCrunch.]

If you haven’t used Powerset’s first announced product, semantic searching of Wikipedia, check it out on their site and you’ll begin to see why there’s been so much interest in their technical approach. I’ve known founder Barney Pell for a while now, and we’ve mused about the possibilities of adding Powerset’s strengths to Microsoft’s global scale.  The more I played with PowerLabs, before its full launch, the more I was convinced of its power.

When I was working at DIA, one of our dreams was a semantically enabled intelligence enterprise. IC analysts and advanced users within any other enterprise vertical are going to find some very interesting capabilities finally possible when Powerset technology is wedded to the FAST search software already being deployed at web scale. But that’s only the beginning.  I knew of Microsoft’s long history in semantic research, dating back to the MindNet days and before. Microsoft Research has been ramping up our investment in semantic approaches for a while now, because we see enormous benefit in embedding semantic capability within the billions of web services and software services we’re planning, in the “software plus services” plan.  Keep in mind:

  1. We’ve already been delving deeply into graph and subgraph relationships among information and documents at web scale (see “Web Projections: Learning from Contextual Subgraph Projections of the Web,” 2007).
  2. We’ve added to that some ambitious partnerships with the brightest minds in academic and commercial semantic research, with our program “Beyond Search: Semantic Computing and Internet Economics,” launched late last year “to improve the ways in which the information seeker finds, shares, discovers information.” (Check out the impressive list of projects at the link.)
  3. Semantic understanding will only enhance the big vision – and Craig Mundie describes that this way: “We tend to believe that there will continue to accrue a large amount of computational capability in the literally billions of intelligent gadgets and devices.”  Software on those billions of devices will coordinate and interact with a web-scale “services layer” of cloud-supplied intelligence.  One avenue of that activity is imagined in last year’s Microsoft Research-sponsored paper on “Semantic Reality: Connecting the Real and Virtual World.”  That paper describes SR as integrating “a large body of work in sensor networks, embedded systems, ambient intelligence, networking, distributed systems, distributed information systems, artificial intelligence, software engineering, social networking and collaboration, and Semantic Web.”

Simple but surprisingly powerful examples are already appearing, as in the Microsoft Research “Blogosphere Early Warning System” or BLEWS, which was unveiled in prototype at the 2008 TechFest and may go live during this presidential election. Here’s a screenshot of BLEWS; intelligence analysts can think of it as an advanced OSINT tool, consuming web-scale social media and blog content.  BLEWS screenshot BLEWS semantically identifies political blogposts, ascribing a left-leaning or right-leaning interpretation.  it then extracts and mines all links to news articles, applies an “emotional charge” classifier and a de-duplication algorithm to related articles, and finally produces a visualization of the volume and semantic relationship of the articles in near-real-time.  Nifty! A personal disclosure: my happiness at the Powerset acquisition isn’t just because of the technology it brings to Microsoft. In fact, when I was preparing to leave government service last year, I was approached by the startup about their position of VP of Engineering.  I had some fascinating discussions and love the company, the technology, and Barney Pell – but I felt I was a better fit as a generalist at Microsoft.  Well, what do you know – happy endings!  E pluribus unum 🙂    Email this post to a friend

AddThis Social Bookmark Button

2 Responses

  1. […] Also, I admit it, I can’t stop playing with the semantic-space startups, whether it’s Sensebot or the newly launched Evri, or Powerset (just acquired by Microsoft, as I wrote about previously). […]


  2. […] embed semantic techniques and power. We continue to build upon Powerset technology for example (I first wrote about that back in 2008).  And it was just six months ago – but seems like years – that I wrote […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: