Kngine (pronounced kin-gin, short for knowledge engine) is one of those startups with a goal that’s both straightforward and impressively ambitious — it wants to build an app that can answer any question. In fact, when you open the app, it prompts you to “ask me anything.”

When I watched the promotional video (embedded below), the first thing I thought of was Apple’s Siri. And while Kngine co-founder and CEO Haytham ElFadeel doesn’t shy away from the Siri comparison, he also said Kngine has a slightly different goal. One of Siri’s big selling points is allowing you to access a lot of the iPhone’s functions through voice, so when your questions are more fact-based rather than task-based (i.e. Kngine’s strong point) it relies on Wolfram Alpha.

I haven’t had a chance to give Kngine a thorough test, but when I tried the app out, it was able to answer all of my random questions accurately. (Naturally, I started with “What is TechCrunch?”) ElFadeel also said the company hired an independent consultant to compare Kngine to Siri and Evi in a test based on the NIST guidelines, basically by asking a bunch of different questions. The current version of Kngine answered 54 percent of the test questions (either by delivering the correct answer that showed an understanding of the question, delivering a partial answer, or delivering the correct answer despite misunderstanding the question), compared to 26 percent for Siri and 25 percent for Evi. Among the questions that Kngine could answer but its competitors couldn’t: What band is Fred Durst in? What is the periodicity of Halley’s comet? Who founded the AARP?

Version 2.0 of Kngine, which has yet to be released to the public, did even better, answering 71 percent of the questions.

Behind the scenes, Kngine is constantly crawling the web, not to index pages like Google, but rather to extract knowledge and meaning. ElFadeel compared the technology to Wolfram Alpha, but he said Kngine gathers its data in a much more automated way.

Kngine is based in Cairo, but ElFadeel has moved to the San Francisco Bay Area and is building out a business development team here. The company has raised $275,000 in funding from investors, including Sawari Ventures. The company first launched a prototype in 2010, but it didn’t release a real consumer app until this year.

For now, it’s more focused on acquiring users than making money, but ElFadeel said monetization possibilities include running advertising in the app and also licensing the technology to other companies, say enterprise search products that want a natural language interface.

Article courtesy of TechCrunch

Oh Google. Sometimes you’re so awesome.

Google search has long featured a built-in calculator function but a recent update added a fully functional 34-button scientific calculator. Previously, when a user entered, say, 2+2, Google would simply display the sum above the search result. Now, when that equation is entered into the search bar, the answer pops up along with the new calculator. Best of all, this works in mobile browsers and voice search, too.

This isn’t a stripped down calculator, either. It’s a full-power, voice-enabled scientific calculator with nearly all the functions of a tangible model. Plus, it doesn’t require two AAA batteries. The mobile version lacks the scientific functions and voice control, but there are plenty of scientific calculator apps available. Update: A Google Product Manager pointed out in this post’s comments that the scientific functions appear when the phone is rotated to be viewed in a landscape mode.

This calculator even works with Desktop Voice Search. Simply click the little mic icon and state the equation; it works with both “what is the square root of 30?” and “square root of 30.” Or, to launch the calculator itself, say “calculator”. It seems to stumble on long, complex equations (or maybe I’m saying them wrong), but in the right situation, this voice-powered calculator could be rather valuable.

This is just Google’s latest addition to its nerdy toolbox. The search bar already performed graphing functions. Now, with Google, Wolfram Alpha, and the sheer number of apps out there, there really isn’t any excuse for not being able to finish your math homework.

Article courtesy of TechCrunch

It was written off as a glitch in the Matrix and quickly fixed when Apple’s Siri proclaimed that the Lumia 900 Windows Phone was the best smartphone available. Now, just a few weeks later, Samsung’s S Voice digital assistant points to HTC’s Trophy Windows Phone as the top smartphone. Coincidence? Perhaps. But it’s also possible that the computers are trying to tell us something…

Far from being artificial intelligence, Samsung and Apple’s digital assistants are simply a voice interface to certain web portals. In Siri’s case, she (it?) looked to Wolfram Alpha to decide the top smartphone. Wolfgram Alpha then queued up Best Buy’s user ratings, delivering the result of the Lumia 900 as the best smartphone a few weeks ago. Similarly, Samsung’s S Voice looked directly at Best Buy’s user rating, which now lists the HTC Trophy as the top smartphone (the Lumia is now #4 on the list). Of course these results are a bit tongue in check but there is still a lot of truth here.

Windows Phone is generally considered to be a solid performer. Users love the simple, to-the-point interface. The hardware in the Lumia 900 and HTC Trophy are equally impressive. Siri and S Voice are right: Windows Phones are serious contenders in today’s smartphone wars.

Article courtesy of TechCrunch

Wolfram Alpha, the “computational knowledge engine,” just launched a desktop app for Windows 7 PCs through Intel’s AppUp store. After launching numerous iOS and Android apps (including for the B&N Nook and Amazon’s Kindle Fire), this is the first time Wolfram Alpha is introducing a dedicated desktop app for its service. The application costs $2.99 and is only available through Intel’s store.

There are a few oddities here. First of all, it’s not clear why you would really need a dedicated desktop app for Wolfram Alpha. The app, after all, doesn’t really offer anything that the free web-based service doesn’t also offer. According to the product page, the main selling point here seems to be that you don’t need a browser to use it. For most people, using a browser isn’t much of a hassle these days. The only semi-compelling reasons to use the desktop app are that it comes with “a specialized keyboard” (that’s a software keyboard, of course, not a hardware one) and “extended copy and paste support.”

The other oddity is that Wolfram Alpha is exclusively selling this app trough Intel’s AppUp store, the company’s app store for PCs and MeeGo-based devices. We weren’t able to track down any user numbers for AppUp, but the fact that even the most popular free app in the store only has one review makes us think that it isn’t a huge hit by any means.

According to today’s announcement, Wolfram Alpha plans to bring all of its mobile apps to AppUp, starting with its Course Assistant apps. There’s no mention of other stores or platforms, so it’s not clear if the company has any plans to bring its own desktop apps to Apple’s Mac app store anytime soon.

Article courtesy of TechCrunch

Locu, the data-focused startup launched out of Sir Tim Berners-Lee’s lab at MIT to provide structure to the world’s information, has just announced a $4 million Series A round led by General Catalyst Partners. Also participating in the round were Chris Sacca’s Lowercase Capital, Lightbank and SV Angel, as well as Locu’s existing angel investors, Naval Ravikant, Babak Nivi, Quotidian Ventures, and Matt Ocko of Data Collective.

The new round follows Locu’s raise of over $600,000 in seed funding back in September 2011.

If you haven’t heard much about Locu yet, you wouldn’t be alone. Co-founder Rene Reinsberg says the company has been “flying under the radar” and has just been focusing on building their product.

That product is technology to structure the world’s information – a refrain that sounds a lot like a follow-on to Google’s goal of organizing the world’s information. It’s the next logical step.

Explains Reinsberg of what his company does, “if you look at the ways people are currently making sense of unstructured information, it’s often an approach that’s based on building crawlers and maybe some machine learning techniques – that works well if the data is out there in an already semi-structured format,” says Reinsberg. “But oftentimes, you come across really messy, distributed data sets for which automation or any machine-based approach doesn’t work, so the only way you can really tackle it is manual labor and the brute force approach.”

“So what we’ve built,” says Reinsberg, “is a system that leverages both the machine learning side and crowdworkers.”

But Locu’s crowdworkers are not Amazon Mechanical Turk users…they’re skilled, technical types who are trained (for around two hours) in how to use Locu’s special markup language to map data to the schema. Before they’re set loose on the job, they’re even quizzed first to ensure they’ve understood the system properly.

Their job, then, is to clean up, correct, tweak and otherwise adjust the data the machine-learning side of the system hasn’t managed to translate 100% correctly. Locu has, to date, trained thousands of these workers, but only a couple hundred or so are actively working to make sense of the data at any given time.

For now, the first vertical Locu has been attacking with its proprietary technology is local business information, and specifically, restaurant menu data – something that Reinsberg says is “still a really big pain” for the industry.

Locu already has some early customers and traction, but Reinsberg can’t disclose who it’s working with right now. The customers are really big companies, which you would have heard of, he says, but Locu is also working with a few early stage startups as well.

Going forward, the startup will use its system to build up a database of other non-restaurant business data, like the price lists at personal services businesses (spas, hair salons, etc.), for example. But further down the road, the technology could be used to provide structure to other data sets, like patents or healthcare. Locu will also look into allowing enterprise to use its technology to structure any data they would have a need for, whether internally or externally sourced.

The timeliness of the system is ideal for the smartphone age, Reinsberg believes. “We think the platform that we’ve built is so important is because more applications rely on structured data,” says Reinsberg. “Search is moving towards more towards structured data. You look at things like Siri and on the backend, Wolfram Alpha…and you can see how there’s much more emphasis going forward on having structured data in a variety of different verticals,” he says. “Local for us was an obvious one because of the rise of smartphones. People want to have a lot more actionable data at their fingertips and in the apps that they’re using,” he adds.

Locu, which currently houses its team of thirteen in Cambridge, is now opening an office in San Francisco, and will be using the new funding for product development and hiring.

Article courtesy of TechCrunch

Looks like Apple might be loosening its grip even more on voice recognition apps? Or, it simply just feels that the competition is not as good as its own native Siri. We’ve just gotten word from Netherlands-based developer Sparkling Apps that its voice-response app, Voice Answer — rejected by Apple for the nearly three months — has been approved by Apple and is now live in the App Store, and usable on any iPhone, iPod or iPad running iOS 4.2 or later.

It took “almost three months of negotiating, tweaking and pushing,” developer Martijn van der Spek tells TechCrunch. Like Siri, the app is based on data from Wolfram Alpha, among other sources, and lets users ask questions by either speaking to the app or typing in a question. It’s priced at £2.49 ($3.99).

He says the company is now going “full speed ahead” implementing more features into the app. These include location-based place finding and email/SMS and more voice function commands. Additionally it’s adding in a bit of sci-fi kitsch: it’s planning to create an animated robot for the interface. You can see the video of how that will lookbelow.

The news comes on the back of other voice recognition apps making a splash and then facing rejection issues with Apple, perhaps most notably Evi.

Sparkling Apps in March had a free voice recognition app, Talk to Eve, also rejected for being “too similar to Siri” that was subsequently approved in March.

With the voice-recognition space currently very active right now, the big question is whether any of these third-party developers will be able to gain traction against Apple, and what they will all do next to make themselves relevant and indispensable to users. Offering APIs to other app developers could be one lucrative route.

Article courtesy of TechCrunch

Traveling over seas? Bringing your phone?

You’ve got at least two things to remember:

• Put a password on it, and take off all the personal stuff (read: those drunken nudie pics? Yeah, they should probably go.) Getting your phone back after its been misplaced or pick-pocketed is a lot harder when you’re on the roam.
• Figure out how you’re handling your international data, or you’re gonna come home to the nice surprise of a 70 billion dollar phone bill.

Looking to make their International Data plans a bit more enticing, AT&T has just bumped the allotment of data for each plan. Unfortunately, any of the reasonably big allotments still cost an arm and a leg. Hell, more like multiple arms and legs. I don’t know, how much do arms and legs go for these days? Can I use Wolfram Alpha for this?

The new International Data Plans:

• $24.99 gets you 50MB (used to get you 20MB)
• $49.99 gets you 125MB (used to get you 50MB)
• $99.99 gets you 275MB (used to get you 100MB)
• $199.99 gets you 800MB (used to get you 200MB)
• Every 10 megabytes you use over your plan will cost you $10

At over double the value of the old plans, the new plans are certainly an improvement — but you’d have to be pretty crazy to use them. Pro tip: in almost all tourist-friendly countries, you can rent a SIM with unlimited data for a few bucks a day.

Article courtesy of TechCrunch

Wolfram Research, the parent company behind computational search engine Wolfram Alpha, is acquiring MathCore Engineering AB, the developer of the MathModelica modeling and simulation software system. Terms of the deal were not disclosed.

MathModelica is a software system for multi-engineering modeling and simulation based on Modelica and Mathematica. MathModelica allows users to develop advanced multi-engineering models in a simple drag and drop interface and includes a large number of built-in base models for engineering domains, including electrical, mechanical, and thermal, as well as for areas like biochemical modeling. The resulting models can be directly simulated and visualized. MathCore’s software is used by companies such as Rolls-Royce, Siemens and Scania.

One of the pluses of MathModelica is that it is fully integrated with Wolfram’s computational software program Mathmatica. MathModelica models can currently be directly imported into Mathematica. And the two software products are complimentary. For example, measurement data can be imported to Mathematica and used to validate models, identify parameter values, and so on.

Wolfram plans to bring together Mathematica and MathCore’s technology—as well as Wolfram|Alpha and CDF—to create a system that “will launch a whole new era in design, modeling and systems engineering.”

For more details on the acquisition and technology, checkout Stephen Wolfram’s in-depth blog post on the subject.

Article courtesy of TechCrunch

The recent victory of IBM’s Watson computer against human competitors in an exhibition round of Jeopardy got computer scientist Stephen Wolfram thinking about how regular search engines might fare in such a match-up. So he took 200,000 known Jeapardy clues and ran them through six search engines (Google, Bing, Ask, Blekko, Wikipedia Search, and Yandex). He excluded known Jeopardy sites from the results, and didn’t test his own Wolfram Alpha because it is not designed for those kinds of queries.

What he found is that the search engines did fairly well, depending on how you measure success. Google did slightly better than the rest, but Bing and Ask were close behind. On average, Google got the correct answer somewhere on its first results page 69 percent of the time, versus 68 percent for Ask and 63 percent for Bing. Google got the right answer somewhere in the title or snippet of text of the very top result 66 percent of the time, versus 65 percent for Bing (and Ask dropped to 51 percent).

In comparison, most humans answer 60 percent of Jeopardy clues correctly, while the top player of all time, Ken Jennings, answered 79 percent correctly. So it is conceivable that a system could be created using regular search engines that could beat most humans. But Wolfram cautions:

Of course, the approach here isn’t really solving the complete Jeopardy problem: it’s only giving pages on which the answer should appear, not giving specific actual answers. One can try various simple strategies for going further. Like getting the answer from the title of the first hit—which with the top search engines actually does succeed about 20% of the time.

Answering Jeopardy clues correctly and consistently is a hard problem for computers to solve because of all the variations and nuances of human language. Yet “just using a plain old search engine gets surprisingly far,” concludes Wolfram.

Article courtesy of TechCrunch

Computational search engine Wolfram Alpha is debuting its Android App in the Android Market today, simultaneously launching with the new T-Mobile G2 Android phone.

Similar to the iPhone app, WolframAlpha for Android is available for $1.99 in the Android Market. The app features the ability to search with a voice keyword, includes ten trillion data elements of knowledge, and tens of thousands of computational models.

WolframAlpha’s allure is the engine’s ability to provide you detailed numerical data on a vast variety of subjects including math, science, engineering, health and nutrition, geography, economics, linguistics, people and history, sports, and music. For example, you can enter a math problem into the search engine, and it will return with the correct answer. Or you can type in a type of food, and the search engine will return the nutritional data.

It’s probably a good idea that the search engine avoided pricing its Android App at $50 as it did initially with its iPhone app (which was eventually slashed from $50 to $1.99 after less than stellar downloads). This more moderately priced app on Android phones should bring a faster rate of adoption.

Article courtesy of TechCrunch