WebProNews » Powerset

Powerset Finds Its Way Into Live Search

Chris Crum — Thu, 18 Sep 2008 00:01:07 +0000

Powerset, a search engine, which was once touted as a "Google Killer", and was sold to Microsoft for $100 million a couple months ago, is now being integrated into Microsoft’s Live Search results (at least to a limited extent).

Not many users will actually see these results yet as the Powerset Blog notes. They are simply running some tests. After some testing projects, they have big plans to use Powerset’s technology and Live search’s combined to feed each other, presumably to achieve the most relevant search results possible.

Test 1 – Expand the number of queries for which Live Search shows Answers.

Test 2 – Use Powerset’s semantic technology to generate improved captions for Wikipedia articles.

Test 3 – Use Powerset’s Factz extraction to generate a list of related searches for a set of queries.

Unfortunately, I am not one of the few "lucky" ones to be able to see the tests, but the Powerset Blog provides some screenshots of results. For a demonstration of what Powerset does, check out this video:

Powerset Demo Video from officialpowerset on Vimeo.

Will Microsoft make up some of that search market share with Powerset? Maybe, but I don’t expect them to make a huge dent in Google’s. Like Jason said before, it’s Google’s brand that it hasn’t been able to overcome, and that brand is stronger than ever. Google Killer? I doubt it.

Report: Bill Gates Remains Tied To Search

Doug Caverly — Wed, 20 Aug 2008 21:10:01 +0000

The latest Nielsen stats put Microsoft’s share of the search market at an unimpressive 11.9 percent. What’s more (or to be accurate, less), its year-over-year growth is negative. But Microsoft’s search team has an interesting ally in its corner, as it turns out Bill Gates may still be lending some sort of hand.

Gates was supposed to have stopped participating in Microsoft’s everyday business about two months ago. Philanthropic pursuits would more or less occupy the rest of his life.

Only Stephen Shankland reports that Scott Prevost, the general manager and product director at Powerset, said at SES San Jose, "Bill has definitely not retired for us."

Having Gates on hand should provide Microsoft’s search team with a huge morale boost, or at least a motivational dollop of fear. There is the question of why Gates would choose to do this, yet some answers are obvious (he’s interested in search, he doesn’t like losing) and honestly, a billionaire’s thought pattern might not be easily followed.

The one other thing to consider is whether Prevost is simply wrong; Powerset represents a relatively fresh Microsoft acquisition, so perhaps the newbie failed to recognize the age of some old memo.

It’s likely that a life of writing checks day and night wouldn’t agree with Gates, though, and Microsoft’s search engine may be getting a little personal attention as a result.

Powerset: Why A Google-Killer Sells So Cheap

WebProNews Staff — Thu, 03 Jul 2008 21:58:42 +0000

Of all the post-Google search engine startups, Powerset was touted as the most likely Google-killer. We waited with bated breath, we rubbered our necks around to get a better glimpse—early reports were stellar, yes this was the one to take on Google and. . .Wait, what?

Sold to Microsoft. $100 million? Wait, what?

A Google-killer’s worth more than that right? How many billions is Microsoft still thinking about shelling out for Yahoo’s search engine alone, which ran neck and neck with Google for exactly five minutes? And heck, that wasn’t always even because of difference in relevance. Studies came out saying Yahoo’s and Google’s results were pretty much identical—it was that Google brand that couldn’t be overcome.

If a plucky upstart could possibly outdo Google and establish a killer brand, isn’t that worth more than $100 million? Why the swift sellout?

The hard-nose skeptic throws that proboscis in the air and intakes the sharp smell of bovine repast…all hype, no delivery. But still, $100 million is a good payday for Barney Pell, and Microsoft doesn’t make a habit of bidding on search engines unless they’re Yahoo stature—Steve Ballmer’s need to fill his Internet hole aside. So what’s in it for both of them?

Don Dodge, Director of Business Development for Microsoft’s Emerging Business Team, blogs that Powerset didn’t have much choice but to sell to somebody with much deeper pockets. The technology was powerful, even game-changing, operating not on the query side but on the contextual indexing side—a search engine that can read and understand, basically. . .or not so basically.

And not so cheap in money, memory, infrastructure, or scaling, either. Things as they are, Om Malik reports, Powerset takes about a full second for every sentence indexed. That’s not fast enough to index the continuous exponential growth of sentences on the Internet—How many sentences, you think, have been born since you started this article?

"As you can imagine this is a huge scaling problem, that has been impossible to solve economically until now," writes Dodge. "With Moore’s Law applied to constantly reducing the cost of computing, storage, and bandwidth, it is now possible to solve this problem, although it is still very expensive."

And who has lots of money? Big Daddy Microsoft, that’s who. Who has a burning desire to kick the effing ess out of Google? Same Who. Who didn’t have the money to make that happen and was stuck in neutral until they found it?

Yeah, Powerset.

Microsoft likes Powerset for other reasons, too. Dodge outlines what the technology could do for enterprise search and targeted advertising, both multi-billion dollar markets, and vertical search as well.

It’s hard to feel sorry for somebody walking away with nine digits worth of Microsoft’s money—ah, tropical wishes and Ferrari dreams—but you do get the feeling he got the short end of the same stick Ballmer tried to poke Yahoo with. Gotta hand it to Mr. Softy for those leverage buyouts.

Microsoft Confirms Powerset Acquisition

Doug Caverly — Tue, 01 Jul 2008 20:55:38 +0000

Satya Nadella, the senior vice president of Microsoft’s Search, Portal & Advertising Platform Group, isn’t afraid to admit that the current crop of search engines suffer from some problems. Nadella seems to be hoping, however, that his company’s acquisition of Powerset will help it solve them.

Satya Nadella

Yes, Microsoft and Powerset are fessing up to the deal that was discovered last week. Perhaps since the younger company’s abilities remain somewhat unproven, Nadella writes on the Live Search blog, "We’re buying Powerset first and foremost because we’re impressed with the people there."

He also didn’t get into the issue of price, which leaves us with the $100 million estimate from earlier.

Yet natural language search will be quite important to Microsoft as Nadella later mentioned a "shared vision . . . to take Search to the next level by adding understanding of the intent and meaning behind the words in searches and webpages." And a sizable portion of Microsoft’s resources should be brought to bear on developing Powerset’s technology.

Powerset’s team and headquarters will remain intact in San Francisco.

No Surprise: Powerset Sells Out To Microsoft

WebProNews Staff — Thu, 26 Jun 2008 23:39:55 +0000

A reported price tag of $100 million will deliver the much-hyped Powerset and its natural language search technology to Microsoft.

Matt Marshall at VentureBeat said he’s learned Powerset will go to Microsoft in a deal to be announced in July. The price tag should top $100 million.

Powerset has been something of a media darling through its development. When the site finally debuted, it arrived as a Wikipedia search engine. “I beg that you forget you ever heard ‘natural language’ being associated with Powerset. That’s not really describing what they do in comparison to regular search engines,” Danny Sullivan said in May after it launched.

So Powerset turned into more of an “understanding engine,” able to automatically provide something of a summary of a topic on Wikipedia, it’s not much different than what Wikipedia’s human editors already do on an ongoing basis, with results one can find already in major search engines.

Back in September 2007, when the buzz reached a new level for Powerset, we were convinced it was destined for a payday rather than a space as a stand-alone competitor:

If there

Powerset Follows Hype With Debut

WebProNews Staff — Mon, 12 May 2008 12:58:13 +0000

After a lot of blog buzz, Powerset finally launched its search engine with the initial aim at pulling information out of Wikipedia more effectively.

Ask a question, and Powerset should have an answer, via the online encyclopedia, Wikipedia, as well as an open, community created database called Freebase. The search site delves into the entries to pull out what it calls Factz about a query.

These Factz come back as a simple sentence, derived from what Powerset finds with its query; they appear as subject, relation, and object, according to the Powerset FAQ. “Factz do not always represent truth, but rather propositions that are asserted in the text of Wikipedia,” it said.

As Powerset isn’t searching beyond Wikipedia and Freebase yet, it’s too soon to consider whether or not their approach presents a challenge to the typical contextual search engine market dominated by Google.

At launch, Powerset impressed Danny Sullivan with the way the site “adds some nice value to Wikipedia.” Powerset’s ability to determine relationships across pages in Wikipedia helps it pull out some otherwise obscure points for further refinement of a query.

Powerset faces the problem of pulling in an audience and building a userbase, as all new applications do. Sullivan noted the heavy influence search engines have on driving traffic to Wikipedia, something that Powerset probably won’t receive in the same volume.

The real evolution for Powerset will come if and when it scales its index to handle web search. That’s no easy task, especially considering the billions of pages indexed already by the Googles and Yahoos of the world.

Scale doesn’t come cheap, which to us reinforced the idea we’ve always had about Powerset. We won’t be surprised if they sell out long before the site ever develops beyond what it is today.

Wikipedia Defers To Google

WebProNews Staff — Tue, 08 Apr 2008 21:26:53 +0000

UPDATE 4/9/08: Seems there was some confusion sparked by this piece, which is, at best, a muddled, meandering, word-labyrinth posing as an unconventional, smart-alecky essay on web journalism/blogging. That’s okay, any reader would find himself in good company lost amid my verbose effluvia. For crying out loud, I used a word like "bildungsroman" and made references not just to Roman mythology but also to an obscure Persian king 3,500 years dead. I take full responsibility for that and for trying to force readers to read between the lines to understand this piece wasn’t really about Wikipedia or Powerset. I had an opportunity to illustrate a point rather than just come out and say it, and I took it, and if more than one person doesn’t get it then I failed at illustrating it effectively. C’est la vie. I’ll try to refrain from being too artistic with my points in the future.

So, let’s make the point clear: Online journalists (and increasingly traditional journalists) and bloggers, such as the one described in a New York Times piece whose ambition leads them to insomnia and cardiac arrest, are under enormous pressure to not just be first with a story, but also to evaluate and analyze first, and to do both things more often. This trend makes it more difficult for writers to do their job effectively and responsibly. They have less time to fact check, more time speculate, and incentive to get that speculation out there, which only serves to cloud the truth, which is the ultimate goal of journalism.

It’s also hard on the PR industry because suddenly there are all these new sources and writers to work with, and the Internet just increases the number of possible requests for comment that come in, and all of them can’t possibly be answered. They have to judge who is most important to answer. Luckily, Wikipedia thought I was one of those important people to answer and I was able to dig up the truth of the matter, which is presented before going on to what I thought would illustrate the point via what I thought was quite obviously a fun bit of "faction" (fact + fiction), the point being that the truth is often much different than what can be imagined (but you probably already knew that). Powerset’s Mark Johnson notes also that it is difficult for natural language search engines to separate the two. That doesn’t surprise me. That’s like asking a robot to understand sarcasm and the meaning within voice intonation. Sometimes a human can’t convey that right to another human, much less a machine.

So, below is yesterday’s essay. I’ve reformatted and put the important parts in bold, such as the quote from Brian Vibber, the thesis (which comes awkwardly toward the middle), and the subtle grammatical clues beginning with "just imagine if," which indicate in grammar that what follows is presented in the subjunctive mood, or as we might know it, idle speculation. The speculative part has been put in italics to separate it.

I even added open and close imagination tags.

Perhaps next time, I’ll publish with a disclaimer, an idea that, frankly, kind of saddens me.

Blame information overload. If you did a search on Wikipedia today, you might have been greeted with this message: "Wikipedia search is disabled for performance reasons. You can search via Google or Yahoo! in the meantime."

That got me all excited about the possibilities (cuz search is in my "beat") and immediately sent me into a fit of speculation via keyboard. I asked questions nobody answered; I took screenshots; I relayed a humorous egg-head anecdote, a sort of flash-literary bildungsroman about how I discovered it, and by the time I had finished, well, my question was answered, so let’s save some time and get it out of the way, in a more blogger fashion:

Wikipedia took down its search engine today for maintenance and let Google and Yahoo conduct searches for them instead. Wikipedia search sucked before. Now it might be better because searchers can choose between MediaWiki, Google, Yahoo, Windows Live, Wikiwix, and Exalead. You may have been able to do this before. I don’t remember because I tried to search there once or twice and said never again. This time, though, Wikipedia was pretty good at bringing back relevant results for my favorite imposter king, Smerdis, also known as Gaumata.

Wikimedia Foundation CTO Brian Vibber said, "Search was temporarily disabled as a load-reducing measure during the investigation of an unrelated problem (a change to log page lookups which used bad indexing, bogging down the database servers). It was reenabled a few hours later, once the unrelated problem had been fixed."

Hmmph. A lot can change over lunch.

A few things remain true, though. Powerset and Wikia Search are still out there somewhere under wraps and ambitious rhetoric, and Wales & Co., after all that jazz about Wikia Search, made Wikipedia search better by deferring to the experts. Props for making it better, though. I love Wikipedia.

But it also kind of illustrates the problems with this century’s great media transformation as writers and bloggers are expected more and more to sacrifice thoroughness for speed. (Which seems more important now that we know bloggers are killing themselves trying to be first )

Too bad, though. Just imagine all the speculative buzz that could have been generated if I had published this:

I stumbled across it by typical stream-of-web-consciousness accident, in case you really think I have time to test every prominent site’s search functionality. Previously, during my morning research routine, a story out of India about a baby born with two faces had caught my attention. A colleague joked:

"One could name the child Janeus I suppose."

I told my colleague that was likely very funny in a sad, dark humor kind of way, but I was too pathetically educated to get it. This must be what talking to Dennis Miller is like.

He sent me a link to Wikipedia, which, in his instant-messaging haste was absent a URL parameter. This led me to the Wikipedia search page (and later to the conclusion that, yes, naming it Janeus would be darkly, horrifyingly hilarious and awful—this is the same guy who introduced me to Goatse, after all). I tested it with another query, this time from my own library of obscure references, for King Smerdis, and sure enough got the same message.

I’m not going to pretend that it was a secret Wikipedia’s search engine was severely lacking. I learned a long time ago it was faster and more relevant to run a site-specific search on Google. So it made perfect sense that they would upgrade it eventually (and hopefully).

But I also remember those grandiose predictions made by Wikipedia founder Jimmy Wales about how his search project Wikia Search would bust Google’s block. The mind goes wild with possibilities:

1. Wikia Search is about to launch after just one year in development and they’re going to show it off by integrating it into one of the biggest sites on the Web. Smart move!

2. But they defaulted to Google (and Yahoo, we presume to avoid the appearance of favoritism—it’s okay, we get it) during the interim, which could mean all that bloviating came to naught along with an admission that search ain’t as easy as one might think it is.

3. Powerset also has talked a big game about the future of search and the company’s semantic approach, even if launch seems repeatedly delayed. We were supposed to hear from them in March, which came and went without much Powerset fanfare. Back in September, when Powerset was still teasing everybody, the Powerset blog mentions Wikipedia for explanatory purposes, but rumors circulating in January also spoke of a grand collaboration between the two.

Wow. All that from a two-faced baby.

Which is it? Well, Wikimedia nor Powerset got back with me in time for publication, so we’ll go on wondering for a while.

Powerset Exec Shuffle Looks Like Comedy Sketch

Doug Caverly — Fri, 02 Nov 2007 21:19:17 +0000

In “Monty Python and the Holy Grail,” the Black Knight always gets a laugh by claiming that the loss of both his arms is “just a flesh wound.” Now Powerset’s COO has left the company, its CEO is changing roles, and it, too, claims to be okay.

Raise an eyebrow, snicker, or do whatever else you like, but Monty Python fans will remember that the Black Knight survives his encounter with King Arthur, and so perhaps there’s something to Powerset’s assertion. Heck, Barney Pell, who’s trading titles from CEO to CTO, seems downright happy about the shakeup.

“The changes we are making now position us for a next phase that promises to be really exciting,” he writes near the end of a 1,000-word blog post. “We will bring our technology out in real products that users will enjoy and that will trigger changes across the entire ecosystem of search. I think the next year is going to be an amazing time for Powerset and I am as passionate as ever about Powerset, our technology, our team and our future.”

All right. We’ve heard some implausibly positive stuff out of Powerset before, and not much came of it. But Steve Newcomb, the guy who said that stuff and announced a September release, is actually the now-gone COO, so it could be time to give the company another chance.

Or, if things go poorly, time to watch Powerset bleed on Google.

Avatar Seeks Semantic Search

WebProNews Staff — Fri, 27 Jul 2007 17:40:26 +0000

Researchers at IBM Almaden have been developing a semantic search process that can delve into unstructured text to retrieve structured information.

Avatar Seeks Semantic Search

While a lot of attention has been heaped upon Powerset and its almost-here natural language search, IBM has been working on a similar technology that may or may not be as close to public debut.

IBM calls their effort Avatar Semantic Search. Right now it doesn’t even have the nice minimalist home page Powerset has for early peek signups, but since everyone’s done reading ‘Harry Potter and the Deathly Hallows’, a little text to read is a good thing.

“Ongoing research in Avatar is at the cusp of a number of disciplines ranging from search and information retrieval to machine learning, information extraction, and probabilistic databases,” IBM announced on the project’s page.

We’ve looked at earlier IBM efforts to pull information out of unstructured resources. Their UIMA developments now occupy a place in the freely available IBM Omnifind Yahoo Edition enterprise search product, for example.

But UIMA is so 2005. While Powerset has drawn upon research performed by the Palo Alto Research Center, aka PARC, IBM reached out to the academic community to complement Avatar’s internal team.

They have approached the semantic search issue in three ways. Developing an information extraction system will allow Avatar to plunge into mounds of raw text, and emerge with structured data based on rules-based annotators.

IBM claimed this extraction system will permit unsophisticated users to build an annotator with Avatar and pull out the desired information from email, web pages, business reports, etc.

Through semantic search, the researchers think they can interpret queries people make, and model the real intent behind a query.

The real challenge comes from an effort they refer to as managing uncertainty and probabilistic databases. They’ve stepped deeply into theory here, well beyond any help Douglas Adams can provide for me.

IBM built momentum with UIMA starting well before I’d interviewed Marc Andrews about it in December 2005. It led to the co-branded, freely available Omnifind product I mentioned earlier, and I have to think Avatar may be on a similar track today.

Powerset: The Robot Ninja Search Engine

Doug Caverly — Fri, 29 Jun 2007 19:04:17 +0000

Powerset, a company that focuses on “natural language search,” recently discussed its to-be-released-in-September search engine, and there’s some very interesting stuff going on. Very interesting, and very hard to understand.

If you imagine a young boy trying to come up with a movie plot – he’d probably suggest something involving a time-traveling robot ninja – you’ll get the right idea. Powerset’s COO, Steve Newcomb, likewise seems to have thrown out every “cool” concept that came to mind.

“Imagine a mashup between Facebook, Digg and Google Apps, but you get to participate in the building of the products that sit on top of our platform,” said Newcomb, as reported by ZDNet’s Dan Farber. “You log into a social network, like you would Facebook, and you get certified to be a Powerlabber. Once certified you can join different interest groups, such as travel, and participate in idea and mashup competitions. QA is embedded and it’s all bloggable.”

I’m baffled, and I’m not alone. “What does that mean?” asks Techdirt’s Mike Masnick. “I’ve read it many times and I still can’t figure it out. He goes on to mention MySpace, Second Life and Wikipedia, of course. It sounds like the company is trying to build the ultimate web platform – which is a good strategy, but it needs to get away from buzzwords and patents and actually explain what makes it useful.”

To be fair, though (not that Masnick isn’t), we’ve seen some impressive stuff from Powerset in the past. Some onlookers have even labeled it a “potential Google-killer.” Yet it looks like we’ll have to wait to hear from those Powerlabbers – or perhaps even hang out until that September release – to find out what’s really going on with at Powerset.