SkyPhrase is a very early-phase company that, according to its web site, has “made breakthroughs in algorithms that enable computers to understand more complex language with greater precision than has ever been possible.” The goal is to improve search functionality but also to give developers a new, easy way to incorporate natural language processing into their apps. The company was founded by Rensselaer Polytechnic Institute Professor Nick Cassimatis.

In January, MIT Technology Review reporter Rachel Metz covered the company and actually reviewed an early version of the technology as applied to searching through tweets and emails. It wasn’t yet trained to do what she wanted with tweets but, she wrote, did a “decent” job searching through emails. Part of what makes it work appears to be its ability to understand conjunctions, even if it doesn’t yet have semantic capabilities: “I could search for, say, ‘e-mails from Bob Loblaw in December and January about recipes with a PDF,’ or ‘e-mails from Bob Loblaw or Tobias Funke about cookies in December,’” Metz explained.

Nanostraws in a cell

Breakout Labs’ other new investment, Stealth Biosciences, is a team of Stanford professors, executives and entrepreneurs that has invented a way to get materials into and out of individual cells and to monitor their activity via electric probe. Called Nanostraws and Stealth Electrodes, respectively, the companies two techniques do just what they sound like they do: NanoStraws let doctors inject or extract material from cells in the aims of advancing research and delivering personalized medicine, while the electrodes “automate long-term intracellular electrical recordings of neurons and heart cells.”

Stealth Biosciences, in particular, seems like a heady endeavor, but that’s exactly what Breakout Labs is all about. Launched in 2011, the organization aims to fund projects too early in their lives to attract traditional venture capital. Those funded aren’t giving up large equity stakes in their companies, but are expected to provide a “modest portion” of their revenues back into the program to fund the next generation of Breakout Labs investments. Other investments thus far include Modern meadow — a company trying to create artificial meat using 3-D printers — and AVEtec, a Canadian startup trying to harness the power of tornadoes for good.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Sector RoadMap: Social customer service in 2013
• How emerging technologies will influence collaboration
• NewNet Q4: Platform mania and social commerce shakeout

The launch is timely, considering Punchfork is shutting down its API at the end of the month after it was bought by Pinterest. Several sites and apps tap Punchfork’s recipe content and search capabilities – for instance, Punchfork powered Evernote Food’s Explore Recipes feature – so it will soon be looking for an alternative.

Yummly’s API, though, isn’t just a Punchfork clone, said Brian Witlin, the search portal’s new head of platform and mobile. Punchfork aggregated content from member food blogs and organized its recipes on social principles. Yummly on the other hand delves deep into the ingredients, cooking methods and the science behind each of the recipes it categorizes. It teases nutritional data out of its recipes, and its algorithms can even infer if a particular dish will be spicy, bitter or sweet. Users, for instance, can use Yummly to search specifically for low-fat or gluten-free dish options or find meals guaranteed to blow the socks off even the most jaded spice fiend.

“There are so many ways we can slice and dice the data we have,” Witlin said. “We plan to offer even more options in the next couple of months.” Yummly, however, doesn’t yet have tools to replace the social context Punchfork provides its customers, but Witlin said it’s in the works.

Initially customers most likely will use the Yummly API to provide more generic recipe content and search in their sites and apps. One of Yummly’s early API testers, search engine DuckDuckGo, uses the API to answer specific recipe queries, basically extending Yummly’s search portal onto its own site.

But developers will eventually be able to tap into Yummly’s technology to make their recipe and cooking services smarter. For instance recipe aggregation apps such as Evernote, Paprika and BigOven store recipes scrapped from all over the web, most of them drawn from the same sites Yummly categorizes. Those companies could use Yummly’s API to organize their customers personal recipe boxes into much more useful categories.

Instead of sorting your recipe library by generic soup, salad, meat and poultry labels, you could sort them by calorie level, salt use, level of spiciness or any of hundreds of different categories that aren’t spelled out in the recipes themselves.

Of course, Yummly can only sort the recipes it catalogs so any recipe you enter manually or from a site Yummly doesn’t aggregate won’t benefit from the API. But Witlin said Yummly eventually plans to amp up its recipe parsing technology so it will immediately scan any new recipe it encounters, adding it to its database.  When that happens, there won’t be any recipe Yummly can’t categorize, Witlin said.

Featured image courtesy of Flickr user lilivanili

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Connected consumer first-quarter 2013: Analysis and outlook
• Content monetization: News licensing and syndication still need marketplaces and infrastructure
• Social media in Q1: commerce and discovery dominated

And although there’s a semantic element to it, this is beyond the realm of semantic search. We’re talking about doing a video chat, sending an email or just surfing the web, and seeing relevant content appear before your eyes. Why? Because the web and, heck, even our laptops are so full of information we don’t always know what to look for or have the extra attention to devote to looking for it.

Most recently, I spoke with Christopher Eakins, CEO of a company called Grapple Data that wants to revolutionize desktop search. Presently, the company’s flagship product, called Aikin, is doing something similar to semantic search on the surface. It’s responding to searches with a list of files, emails, contacts or other content — ranked by relevance — that a standard keyword search wouldn’t detect.

He says the product addresses the problem of information workers “being force-fed more than we can chew,” often across applications that don’t interact with each other at all. One might think of Aikin, he said, as a device that records, indexes and keeps track of everything you do on your machine, so you don’t have to remember specific file names, people or even keywords later on. If you have an idea what you’re looking for, it will find that content and then some.

Going forward, though, Eakins hopes Grapple can do away with desktop search altogether, or at least make it less necessary. That’s where the real innovation comes in. He wants to enable what he calls “information gravitation,” where relevant content would start to surface based on the subject of an email someone is typing, for example. It’s like those targeted ads in Gmail, only in real-time and, presumably valuable to users.

I first came across the concept in April 2012 while covering a company called PureDiscovery. Historically dedicated to semantic search and indexing within large corporate datasets, PureDiscovery CEO Dave Copps explained the company’s plans for going much, much bigger. Essentially — first within corporate networks and then across the entire web — it wants to teach is BrainSpace software to learn how people and pieces of content are related and then surface both automatically based on who you follow, what your interests are or even what text you highlight on a web page.

The plan appears to be coming along. The web part, which is definitely a bigger-picture undertaking, seems to have materialized in the form of a beta-mode application called Grokkit. (I’m still waiting for my invite.)

There’s also the work that Expect Labs is doing around its MindMeld application, which my colleague Om Malik lauded as “herald[ing] the era of anticipatory computing.” MindMeld is a video-chat application that also uses voice recognition and some serious data analysis to figure out what a conversation is about and surface relevant information related to that from the web or users’ social graphs. It also tries to predict where a conversation is going and queue up content that it thinks will be relevant in the future.

The point of all of this stuff — and even some of what we’re seeing in the enterprise IT world with startups like Ayasdi and BeyondCore — is that people don’t always know what they’re looking for or the right queries to enter in order to find it. If more information (or at least more relevant information) really is better, this should be a welcome trend.

Feature image courtesy of Shutterstock user photobank.kiev.ua.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Sector RoadMap: Social customer service in 2013
• Connected consumer first-quarter 2013: Analysis and outlook
• Readers weigh in: future prospects for Twitter

Putting food ads on a site devoted to recipes is a no-brainer, but Yummly is bringing some more complex data science to its platform than just keyword search ads. Yummly is applying the same algorithms it uses to parse recipes to its advertising, CEO David Feller said. So instead of popping off ads for boxed pastries whenever a user searches for “apple pie,” the engine can delve deeper.

An apple pie search might generate ads for the necessary ingredients such as for flour or apples, but it wouldn’t just select any variety of apples. Through recipe analysis, Yummly knows that apple pie is most often cooked using tart apples like Granny Smiths, so ads for red delicious apples just won’t do. Yummly can even infer what pairs well with a dish, Feller said.

“The data knows what we don’t know,” he said. “If somebody is looking for apple pie, Yummly knows that apple pie is complemented by ice cream, so we could give them ads for ice cream.”

Though those examples may seem rather simple, Yummly is launching with a limited number of advertisers and ad inventory. But the semantic search engine is growing quickly. In the last six months unique visitors on its site have nearly doubled, growing from 4 million in March to 7.5 million in September. That kind of growth won’t escape the notice of food brands. Feller thinks it won’t be long before Yummly can start delving down into very specific data to hone its ad targeting.

For instance, Yummly doesn’t just parse recipe ingredients. It extracts nutritional data and is even able to determine whether a dish or group of dishes is salty, sweet, savory or spicy from an analysis of its ingredients. Search options combined with histories generated by registered users tell Yummly about an individual cook’s preferences, whether he or she has food allergies, is on a fad diet or favors Asian over Mediterranean cuisines. Feller said all of these factors can be factored into ad targeting, and can be further refined by tracking seasonality of ingredients and even local weather conditions (stews and braises always go over well when its wet or cold).

The associations we make when craving food don’t always make logical sense. If I have a hankering chicken fried steak, it doesn’t mean I’m craving chicken or steak. Instead, I want comfort food, and Yummly would do much better to recommend me mac n’ cheese or meatloaf products and recipes. The ultimate goal, Feller said, is that Yummly will become wired the same ways are brains are to think about food

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Connected consumer first-quarter 2013: Analysis and outlook
• What’s driving the next phase of the e-commerce evolution
• Survey: How apps can solve photo management

A capability like this has been a long time coming considering the problems that exist in searching for patents — especially prior art. Priort art is generally defined as “all information that has been disclosed to the public in any form about an invention before a given date.” For practitioners, however, the definition might as well read “nearly impossible to search for given the ever-growing volumes of published material.”

With that in mind, here’s what Google claims it can do in a blog post explaining the new feature:

With a single click, it searches multiple sources for related content that existed at the time the patent was filed. …

The Prior Art Finder identifies key phrases from the text of the patent, combines them into a search query, and displays relevant results from Google Patents, Google Scholar, Google Books, and the rest of the web.

Additionally, Google blog post author Jon Orwant notes, Google Patents also can translate searches across numerous languages to make it easier to search for similar patents in European countries as well as the United States.

In March, I explained the problems with patent search — especially through the U.S. Patent and Trademark Office — and identified a handful of big data techniques that might be able to help simplify it. What’s interesting, however, is that the USPTO seems to be offloading some of the responsibility for improving the process to Google by helping load its Patents system with the agency’s data.

That presents a curious dilemma for both the USPTO and the patent bar: Google Patents could be highly valuable as a free service using the cutting edge in search techniques to discover patents, but relying on a commercial entity knee deep in patent feuds across a broad range of technologies does open itself to discussions of whether Google’s interests might ultimately influence its results.

Feature image courtesy of Shutterstock user alexskopje.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Social media in Q1: commerce and discovery dominated
• NewNet Q4: Platform mania and social commerce shakeout
• NewNet Q4: Platform mania and social commerce shakeout

The service’s vocabulary and contextual understanding is pretty limited. For instance, if you ask Genius to “find a restaurant” it will pull up the nearest eatery on Google Maps. But if you ask it to find “nearby restaurants” it searches Maps for a joint named “Nearby.” If you’re like me, you’ve probably fiddled with the button a few times and never touched it again, despite its relative convenience on the phone’s faceplate.

But T-Mobile has given Genius a much-needed overhaul, at least on the latest versions of the MyTouch manufactured by Huawei. Nuance Communications, which powers the voice-recognition features on T-Mobile devices, is upgrading Genius’ capabilities and the features of the Genius service, providing a deeper level of natural-language understanding and integrating the service with a much broader array of content sources beyond Maps and Google Search.

For instance, if I were to ask the new Genius for nearby restaurants, it would not only understand my intent but also do its searching on Yelp, pulling up dining options nearby and displaying their rankings and reviews. If I were to change that command to “make a reservation at nearby restaurants,” it would bring me to OpenTable’s website and display eateries in the vicinity that accept online bookings.

If this sounds familiar then you’ve probably used Nuance’s consumer semantic-search app, Dragon Go, available for Android and iOS smartphones. In fact, if you look at the long list of 200 content partners the Genius can access, they’re the same used by Dragon Go. Nuance wouldn’t acknowledge specifically that T-Mobile is white-labeling the semantic-search app, but it’s pretty obvious that’s exactly what it’s doing — and it’s by no means a bad thing. I’m a big fan of Nuance’s intuitive little search app, and being able to access it in fewer steps is a bonus.

The official line, though, is that T-Mobile has basically upgraded its relationship with Nuance to a kind of platinum status. The old Genius tapped into Nuance’s basic speech-recognition APIs, but it had none of the rules-based language-parsing abilities of Nuance’s more-sophisticated offerings. By adding greater contextual understanding and a host of content providers, T-Mobile may be able to turn a pretty lame voice-command feature into something quite useful. So far, though, it’s only available on the MyTouch and MyTouch Q.

Both T-Mobile and Nuance will be represented at GigaOM’s Mobilize conference next month. Nuance CTO Vlad Sejnoha will be speaking on a panel about the future smartphone interface on Sept. 20, while T-Mobile SVP of marketing Brad Duea will discuss the evolution of voice services on Sept. 21.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• CES 2012: a recap and analysis
• Carrier IQ and the continued erosion of operator trust
• The future of mobile: a segment analysis by GigaOM Pro

For Google, it’s all about collecting and analyzing billions of data points to learn what each one really means. With Knowledge Graph, for example, Google uses a “database of more than 500 million real-world people, places and things with 3.5 billion attributes and connections among them.” It’s those connections that are the key, as they’re what make the system smart enough to know what you’re looking for that wouldn’t naturally show up in a standard keyword search.

Although Google hasn’t come out and said so, I’d imagine the Knowledge Graph utilizes Google’s Pregel graph processing engine. Graph processing and databases are catching on in social networks and other large-scale environments because they organize pieces of data by how they’re connected to one another. Those connections are called edges, and they’d keep Knowledge Graph results both informative and focused because the system knows how closely they’re related in any given circumstance.

This example of a personalized interest graph from Gravity Labs illustrates how one might visualize a graph, in this case the connections between a reader’s perceived interests:

Of course, Google has another tool at its disposal, which is the collective wisdom it’s able to glean from billions of searches every day. So, as Singhal wrote when first explaining Knowledge Graph in May, “[W]e can now sometimes help answer your next question before you’ve asked it, because the facts we show are informed by what other people have searched for. For example, the information we show for Tom Cruise answers 37 percent of next queries that people ask about him.”

Google’s other big announcement today is improved voice search on mobile phones, both Android and iOS. Here’s how Singhal describes the new capability:

You just need to tap the microphone icon and ask your question, the same way you’d ask a friend. For example, ask “What movies are playing this weekend?” and you’ll see your words streamed back to you quickly as you speak. Then Google will show you a list of the latest movies in theaters near you, with schedules and even trailers. … When Google can supply a direct answer to your question, you’ll get a spoken response too.

On Monday, a Google Research blog post noted how the company’s work on neural networks — which it famously used to train a system capable of detecting cats and human faces in video streams — is being used to power speech recognition in the Jelly Bean release of Android. Seventeen-year-old Brittany Wenger recently won the Google Science Fair by building an application atop Google App Engine that uses a neural network to help detect breast cancer.

As one might imagine, however, the big challenge for Google, Microsoft , Apple and everyone else trying to provide intelligent but intuitive user experiences is figuring out how to shape high computer science into easily digestible formats on ever-smaller devices. Search would certainly be a more effective tool if everyone could write complex queries directly against a company’s database, but the trick is making products good enough that we don’t have to. It’s boiling years of machine learning, natural-language processing and neural network research into “you ask a question and your phone spits back the right answer.”

Feature image courtesy of Shutterstock user Sebastian Kaulitzki.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• How to use big data to make better business decisions
• Takeaways from the second quarter in cloud and data
• The importance of putting the U and I in visualization

Admittedly, I don’t spend too much time searching academic or professional databases anymore, but this is a novel approach from what I’ve seen. Powered by Atigeo’s software product called xPatterns, the new interface for exploring PubMed presents a hub-and-spoke-like diagram (which it calls “bubbles and sticks”) that viewers can manipulate by adding and subtracting search terms or by searching for related terms. With every click, a users drills down further into the results, although the original map of terms remains.

Atigeo markets xPatterns to a number of industries, from the public sector to advertising, to help them draw better connections between their data, but this use case is particularly cool. That’s because while semantic analysis is already used rather extensively to produce more-relevant search results (or just to proactively present users with content), it’s not every day someone rethinks the process of how we actually navigate search results. Given a little time for experimentation and acclimation, perhaps the xPatterns approach will catch on.

I don’t see why it has to be limited to scholarly databases either. I can see everyday web users wanting to parse through search results on their favorite content sites using a similar approach. Sure, you can trust a site is delivering you exactly what you want to see, but sometimes it might be nice to dig down, see a little of what the system sees to find that needle in the haystack.

Feature image courtesy of Shutterstock user Michelangelus.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Cloud computing infrastructure: 2012 and beyond
• The importance of putting the U and I in visualization
• A near-term outlook for big data

For starters, Nuance is fully implementing into Swype the XT9 predictive text technologies originally developed by Tegic. Nuance has built on Tegic’s algorithms, though, expanding its prognostication abilities beyond just words to the formation of sentences and phrases.

For instance, the words “pit,” “pot” and “put” follow the same trace path on the Swype keyboard, forcing you to select from the right word from a menu. But by analyzing the words that come before or after the entry, Swype now can place the right word by context. It knows to use the word “pit” after avocado and “put” before the phrase “that away.”

The platform’s intelligence goes beyond just common language structure. Nuance says Swype will adapt and customize itself to its user. It learns to predict phrases and word groupings frequently typed by users. For instance, if I enter the words “Grandpa Fitchard” enough times into text messages, Swype will begin automatically filling in the last name every time I type “Grandpa.”

The new version will also learn from all of your missives, not just the ones composed through Swype. If you give it permission, the software will cull through your emails, SMS messages, Twitter feed and Facebook posts, searching for new words to add to your phone’s personal dictionary.

Nuance has already integrated its Dragon Dictation speech technology – and in some cases Dragon Go semantic search — into Swype, but with this latest release of its software it’s fully bridging the gap between spoken and tactile input methods. Dragon’s speech capabilities use the phone’s data connection to connect to Nuance’s language servers, while text recognition remains local to the device. But Nuance is now linking the two, allowing you to generate a new vocabulary of spoken words synced to your personal dictionary.

For instance, if you spoke an uncommon last name, say “Fitchard,” into the phone, Nuance’s servers simply wouldn’t recognize the word. The Swype interface would then prompt you to type the letters of my last name into the keyboard. It then forevermore associates that spoken word with the written text, creating a unique lexicon of recognizable spoken words for each user.

In addition, Swype now has support for 55 languages — both written and spoken — and offers a third input method (fourth if you count simple key tapping): writing out words by finger, which it then turns into text.

Nuance is selling the new Swype release to Android handset and tablet makers as well as making the technology available to other OSes via a software developer’s kit. If you don’t want to wait for an update from your manufacturer, you can try downloading a beta version for Android from the Nuance site.

A warning though: The beta installer doesn’t recognize all phones, and it won’t load on devices in which a manufacturer has preinstalled Swype in the factory. That involves a lot of Samsung and HTC phones. As with many things Android-related these days, you may be at the whim of your carrier or device maker when it comes to getting the latest version of Swype’s software.

Update: Nuance has told me that the company updated its device policies Wednesday when it released the beta to the general public this morning, making it now possible to download the software on devices that already have Swype factory pre-installed. I tried to download the software on my HTC MyTouch again this afternoon, and, true to Nuance’s word, it worked.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Survey: How apps can solve photo management
• Analyzing the wearable computing market
• Who and what to watch in the new era of the living room

Nuance Communications wants you to have a conversation with your car via the cloud. The speech recognition company already powers many of the voice technologies embedded into today’s automobiles, but on Tuesday it unveiled Dragon Drive, which moves beyond simple voice commands into the realm of natural language understanding.

Voice command technologies like Ford’s Sync allow drivers to audibly initiate basic tasks like playing music, control the thermostat or answer an incoming call, but there are limits to what Nuance’s onboard software can do. But by tapping into a 3G or 4G connection, a driver’s command can reach out to Nuance’s cloud-based language servers, which can not only understand a far broader range of commands but interpret those commands and deliver results with much more sophistication, said Mike Thompson EVP and GM of Nuance’s mobile division.

What does that mean exactly? Well, according to Thompson, the long-term result will let drivers have a natural conversation with your car – something akin to talking with KITT, the sentient Trans Am from the 1980s TV series Knight Rider. In the near-term we’ll see many of the same capabilities found in Apple’s Siri personal assistant: an interface that is not only able to perform basic tasks and search for information, but also interpret context and intent.

“We are already making KITT in many ways,” Thompson said. “There is a Siri-like effect sweeping across all product categories. … There’s a huge demand for it in automotive.”

Nuance is starting small. The first Dragon Drive application will be an SMS service, allowing drivers to send a text message to a number or contact in their address books as well as dictate the message itself. That service will start appearing in vehicles on dealer lots this summer, Thompson said. But soon, Nuance will start layering on more functions.

Many of the features available in Nuance’s consumer services, such as Dragon Dictation and it’s voice-powered semantic search app DragonGo, will make their way into the dashboard, but Thompson said Nuance wouldn’t simply be repurposing all of its applications for the car. The focus for Dragon Drive, he said, will be services like vehicle navigation and points of interest that are much more useful to a driver, who isn’t going to be spending much time scouring Wikipedia or shopping for handbags online (hopefully).

Music is a good example of where the power of a cloud-based intelligence will come to bear. For instance, if you were to ask a Dragon Drive-equipped car to play a track not loaded into your car’s console or smartphone, the service could give you plenty of other options besides silence. It could search your cloud-based music collection for the track and stream it through your in-dash stereo. It could offer to create a Pandora station based on the particular song or artist. Or it could ask you if you would like to buy the track or album and download it over the air.

Image courtesy of Flickr user PopCultureGeek.com

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• CES 2012: a recap and analysis
• Why Google Android’s Electric Vehicle Deal With GM Matters
• How emerging technologies will influence collaboration