Tuesday, June 5, 2012

Voice Biometrics for Mobile

Having trouble remembering all of your passwords?

Voice biometrics is here to help.

Nuance has released Dragon ID, a voice biometrics technology that lets users unlock mobile devices and consumer electronics by voice alone. Dragon ID enables manufacturers to secure and personalize access to phones, tablets, apps, and PCs with a simple spoken word or phrase.

According to a recent consumer survey by Nuance, 96 percent of respondents stated they often make mistakes typing their passwords into their mobile phones. With Dragon ID, people can simply create a spoken password or passphrase, offering a secure and convenient way to wake up and access devices. For instance, just saying “Hello Dragon” can securely turn on a device and instantly launch a personalized home screen with favorite apps and content, even on shared devices like tablets or PCs.

“Consumer devices and the experiences they deliver are incredibly personal—and with that comes a need for greater security,” said Michael Thompson, executive vice president and general manager, Nuance Mobile, in a statement. “Nuance’s innovative Dragon ID further humanizes the mobile experience, leveraging a person’s unique voice to secure and access their device, and keeping the content they rely on each day private and personal.”

Nuance’s voice biometrics authentication technology is already used by the world’s leading enterprises, carriers, banks, and government institutions, with over 20 million voiceprints enrolled globally.

Dragon ID is available for consumer device OEMs, initially supporting English, French, German, Italian, Spanish, Chinese, Japanese, and Korean for Android and other platforms, with additional language availability later this year.

via http://www.speechtechnologygroup.com/speech-blog - Having trouble remembering all of your passwords? Voice biometrics is here to help. Nuance has released Dragon ID, a voice biometrics technology that lets users unlock mobile devices and consumer electronics by voice alone. Dragon ID enables manufacturers to secure and personalize access to phones, ...

Sunday, June 3, 2012

Siri voice recognition and the lean back experience

The interesting change in voice recognition is that we don't have to concentrate on which exact words to say anymore, but that we can literally have a conversation with a machine.

Mobile phones have come a long way from the brick-heavy permutations of yesteryear, with their emergency-phone frame and oversized antenna. Back then, phones were considered a necessary supplementary in the rapidly moving world of business – a use that rendered it a somewhat gaudy symbol of opulence.

Now mobile phones are so ubiquitous as to be one of the main pieces of modern technology that transcends social boundaries – and in its most advanced iterations, highlights an eloquent marriage of sound ergonomic design and superior technological competence. Such is its capacity, that an internet-ready phone replete with a megapixel camera and MP3 player no longer awakens even the most impressionable eyebrow.

Understandably, one of the elements of sophisticated design that has been less quick to catch-up has been voice recognition. I can personally remember some time spent testing voice recognition capability on Microsoft Word in the late 90s. Such software was considered more of a playful distraction than a serious practical tool, such was its inability to recognise phrases and the regularity of misunderstandings.

The current version of the iPhone however, has its own built-in voice recognition (Siri), which uses what is referred to as a natural language user interface - essentially a programme that ignores the traditional route of speech recognition (involving identification of keywords), instead focusing on developing a form of language processing that seeks to understand the nature of any question. In essence, this means transforming verbs, phrases and clauses within sentences into triggers for creating and searching for specific data.

There are two really exciting features regarding this technology within the iPhone format. The first is that it is not simply a distractionary plaything. It works. And it does so 70-80% of the time, which shows the remarkable refinement of voice recognition over the last 10-15 years. The second is that Siri gets better over time, by recognising a user’s voice and learning its preferences and tastes.

The online blog TechCrunch gushingly calls Siri a signal that it is “game over” for the normal way of searching and locating information. It may well be over the long term, but there is undoubtedly some way to go before internet browsing and searching becomes a voice-driven experience. However, as the John Battelle quote states in the aforementioned TechCrunch post, “the future of search isn’t search, it’s a conversation with someone we trust.”

The future lean back reading experience framed as a conversation opens up enormous possibilities for users to broaden and expand their knowledge. A future “conversation” may not involve us scouring the web or an app for articles of interest – it may involve a Siri-like companion doing the legwork for us, bringing up a whole host of possible reading (perhaps even from sources unknown to a user), based on what we have been reading about or commented on in the past. The conversation could even stretch to asking a voice-recognition companion to help elaborate on a point/fact/figure of interest within an article, guiding us in the direction of explanations, further reading, allowing us to compare facts and figures or clarifying definitions.

That this is the beginning of voice recognition is not a matter of dispute. The question is, just how far will it be able to go? 

via http://www.speechtechnologygroup.com/speech-blog - The interesting change in voice recognition is that we don't have to concentrate on which exact words to say anymore, but that we can literally have a conversation with a machine. Mobile phones have come a long way from the brick-heavy permutations of yesteryear, with their emergency-phone frame and ...

Nuance fights to maintain edge in speech recognition

Now that speech recognition has taken center stage, this technology enjoys massive R & D investments....

In the latest batch of television ads for Apple’s iPhone 4s, one high-profile Massachusetts resident takes center stage. But a second, perhaps more crucial local player, remains incognito.

Movie star John Malkovich, who lives in Cambridge, lounges in a leather chair and chats with Siri, the virtual personal assistant built into Apple’s latest mobile phone. “Life . . .,” Malkovich says languidly. “Try and be nice to people,” Siri replies. “Avoid eating fat.” The software that translates Malkovich’s words into text so that Siri can serve up an answer is made by Nuance Communications, a Burlington company that goes uncredited in the ads, on Apple’s website, or in any of the company’s public statements.

The Apple campaign, unavoidable on the airwaves these days, is the highest-profile moment that the rather nerdy field of speech recognition has ever had. Yet the company that has grown into the dominant player over the past decade — mainly through a series of smart acquisitions — doesn’t get a chance to bask in the glory.

“We provide technology for certain Apple products,” says Peter Mahoney, Nuance’s chief marketing officer. “That’s all we can say. If you’ve worked with Apple before, you know they can be fairly restrictive when it comes to protecting their brand.” (Apple didn’t return my phone calls.)

Nuance is in the same position as many other successful Massachusetts tech companies: Most of its revenue — $1.4 billion in 2011 — comes from selling its products to other businesses, and yet it yearns to have a more direct relationship with consumers. When I spoke with Mahoney last week, the challenge facing Nuance was much the same as it was when we talked in 2008: “creating a recognizable, valuable consumer brand,” in his words.

You can’t fault Nuance chief executive Paul Ricci for his vision: In 2001, when Nuance was known as ScanSoft, he paid $40 million in a bankruptcy auction to acquire the assets of Lernout & Hauspie, a Belgian company that had bought up many pioneers of speech recognition. A parade of other purchases followed, including many of local companies such as SpeechWorks International, Voice Signal Technologies, and, last December, Vlingo.

Recently, Nuance has demonstrated new technology called Dragon Drive that can be built into cars, allowing drivers to play certain songs or initiate calls by speaking. Samsung’s latest high-end “smart TVs” include voice recognition software from Nuance that allows viewers to turn them on, change channels, and even initiate Skype videoconferences by speaking. And the free Dragon Go app, available for iPhones and Android phones, gives you 80 percent of Siri’s features, without having to buy Apple’s latest phone.

Ask it to play Count Basie, find the closest Panera, or tell you when “The Avengers” movie is playing and it obliges. (Unlike Siri, it can only search the Web — not the calendar or contact information that’s stored on your phone.) In other words, Nuance is staking out a spot in your pocket, your vehicle, and your living room.

In speech recognition, “they skated to where the puck was going,” says Daniel Ives, a research analyst who follows the company for the investment bank FBR & Co. “Both organically and through acquisitions, they’ve built a massive treasure chest of technology and customers. They’re miles and miles ahead of any other competitor.”

While Mahoney, Nuance’s chief marketing officer, acknowledges that the bulk of the company’s revenue will continue to come from working with manufacturers of TVs, mobile devices, and cars, the company wants to keep exploring products that it can offer directly to consumers, like the Dragon Go app. While it’s free and doesn’t carry any advertising, Mahoney says, “If you can make these direct connections to restaurant reservations or buying movie tickets, potentially you can take a referral fee from those transactions.”

A next step for the company, Mahoney says, is software that doesn’t just respond to spoken instructions, but understands emotion and context. If you sound stressed because you’re late for a meeting and stuck in traffic, for instance, it might suggest an alternate route or offer to send a text message to let colleagues know you’re running late.

But as Nuance buys up smaller speech companies and tries to develop more products to deliver directly to consumers, it finds itself competing with two enormous companies that already have relationships with millions of us: Google and Microsoft. Both have been making major investments in speech recognition.

Google has been focused on improving the speech capabilities that are built into its Android operating system, which it gives away free to makers of cellphones and tablets. And, says industry analyst Walt Tetschner of Acton, Microsoft has been licensing its speech recognition technology to new players like nVoq, which is creating a less expensive medical dictation service that competes with Nuance’s. (Amazon.com also quietly acquired a small speech recognition company last year, and could be adding voice capabilities to its Kindle line of tablets and e-books.)

A decade of acquisitions has turned Nuance into one of those “pillar” companies that Massachusetts so desperately needs. It’s publicly traded, employs nearly 1,000 people in the state, and helps set the agenda for its industry. (Who cares if Ricci, the CEO, spends two-thirds of his time in Nuance’s smallish Silicon Valley office, according to a company spokesperson, and just one-third at the Massachusetts headquarters?)

But one of the things Nuance has sought in buying up so many of its rivals is pricing power — the ability to command a high price for its products and services when it negotiates with customers. Now, the company is facing off against two heavyweights with innovation in their DNA, and the ability to undercut Nuance on price. That means Nuance will have to devise new tactics to stay on the leading edge of the speech industry.

On that matter Siri may not be much help.


The Boston Globe by Scott Kirsner
via http://www.speechtechnologygroup.com/speech-blog - Now that speech recognition has taken center stage, this technology enjoys massive R & D investments.... I n the latest batch of television ads for Apple’s iPhone 4s, one high-profile Massachusetts resident takes center stage. But a second, perhaps more crucial local player, remains incognito. Movie ...

Siri is part voice recognition part artificial intelligence

Speech recognition is just part of the technology that makes Siri work

Apple‘s (AAPL) CEO, Tim Cook, spoke on Tuesday night at the All Things D conference about a wide range of subjects.

SAN FRANCISCO, CA - MARCH 07: Apple CEO Tim C...

SAN FRANCISCO, CA - MARCH 07: Apple CEO Tim Cook speaks during an Apple product launch event at Yerba Buena Center for the Arts on March 7, 2012 in San Francisco, California. In the first product release following the death of Steve Jobs, Apple Inc. introduced the third version of the iPad and an updated Apple TV. (Image credit: Getty Images via @daylife)

One area he was pressed on by Walt Mossberg was Siri, Apple’s voice-automated personal assistant that comes as part of the new iPhone 4S.

“When it works, it works really well,” said Mossberg. “But a lot of times, it doesn’t work.”

Mossberg isn’t the only one to criticize Siri.  Henry Blodget of Business Insider has said ”Steve Jobs would have killed Siri months ago” and that ”Frankly, it’s concerning that Apple is still advertising a product as flawed as Siri.” Fortune’s Adam Lashinsky said that some ex-Apple insiders believe Apple employees are embarrassed by Siri.

Tim Cook’s response, which is here, was that:

- “Customers love it.”

- “There’s more that it can do. We have a lot of people working on this. I think you’re going to be pleased with what we’re doing in the next few months.”

- “Siri’s proved to us that people want to relate to the phone in a new way.”

- Cook implied that voice recognition wasn’t what makes Siri special.  What makes Siri special is the Artificial Intelligence.  ”It’s not voice recognition. It’s the understanding.”

- “I’d put it on the profound list.”

- “This is something that people dreamed of for years. And, it’s here.”

- “We’re doubling down on it.”

I’ve been a big believer in Siri since its introduction last October.  This Forbes article, “Why Siri is a Google Killer,” explains my thinking.

If you took one thing from Cook’s entire talk on Tuesday, it’s that this guy’s a good poker player.  He’s very…. controlled.  Every word – as well as every detail in Apple’s supply chain – is measured.

You get the sense that nothing happens by chance at Apple.  Every move is well-planned.

That’s why I think Blodget and others were supremely arrogant by stating that Steve Jobs wouldn’t have approved of Siri’s release.  Jobs was of course still around and involved in green-lighting Siri before last October’s release.

More importantly, these critics fail to understand the long-term potential of Siri and what’s needed to get there from here.

As Cook said, what makes Siri special is her personality, which is to say the AI.

To get to where Apple (and more importantly its customers) want Siri to be, it needs to improve the AI and for that it needs data.  So, Blodget and others are fundamentally off base by thinking Steve Jobs’ answer to Siri’s current shortcomings would be to pull it and let the guys tinker with it for another year or two back in Apple’s labs.

You can’t get it any better with some tinkering by a few guys. Siri wouldn’t improve at all with another year of tinkering.  You need massive amounts of data.  They’re getting that given how popular the 4S is and the Siri usage they’re currently getting.

Apple said it was in “beta” to let the world knows this is a little different from any kind of product release they’d done before.  But the long-term gain of doing this exercise is going to more than compensate for the short-term pain.

And that’s the reason for all the Siri ads we’ve had since it launched last Fall.  And now Apple’s doubling down on those too.  Why? It’s driving more usage and more data.

It sounds like Siri will be profoundly better before its first birthday 6 months from now.  Just imagine was it will be like at age 2 — or age 7.

So, let the critics howl.  Apple was always about doing something for the benefit of the mainstream. Siri is no different.

via http://www.speechtechnologygroup.com/speech-blog - Speech recognition is just part of the technology that makes Siri work Apple ‘s (AAPL) CEO, Tim Cook , spoke on Tuesday night at the All Things D conference about a wide range of subjects. SAN FRANCISCO, CA - MARCH 07: Apple CEO Tim Cook speaks during an Apple product launch event at Yerba Buena Cen ...

Siri Stats [InfoGraphic] | Voice recognition in numbers

Here's a nice graphical overview how people are currently  using Siri voice recognition as a technology.....

Siri Stats [InfoGraphic]


via http://www.speechtechnologygroup.com/speech-blog - Here's a nice graphical overview how people are currently  using Siri voice recognition as a technology..... ...

Saturday, June 2, 2012

Apple’s Siri likes to talk back- little-known speech recognition commands to get the most out of Siri

Here are are a few commands you probably never heard off to get Siri do more for you.....

Who knew phones could have a sense of humour? Siri, Apple's voice-recognition software featured on the new iPhone 4S, has several witty comebacks for various questions.

Who knew phones could have a sense of humour? Siri, Apple’s voice-recognition software featured on the new iPhone 4S, has several witty comebacks for various questions.

Handout/MCT

One of the most talked-about features of Apple’s best-selling iPhone 4S smartphone (from $159.99 on 3-year plan) is Siri, a voice-activated personal assistant, of sorts.

Rather than typing on the 3.5-inch touchscreen to perform tasks like composing an e-mail, browsing the web, playing music, looking up an address or making a calendar appointment,  these instructions can be done faster and more intuitively by simply asking Siri to do these things for you.

As you might’ve seen in the TV commercials starring the likes of Samuel L. Jackson, Zooey Deschanel and John Malkovich, Siri will also respond to your requests in a human-like voice – a soft-spoken female if you’re in North America.

Related: An easy way to find a lost smartphone

You’re probably aware of some of Siri’s basic abilities, as well as “her” well-documented, tongue-in-cheek answers to questions like “Will you marry me?” or “What’s the meaning of life?” or even “Where can I hide a dead body?” Unfortunately, you might also be aware there are a couple of navigation features not yet available in Canada.

But you might not know of Siri’s lesser-known yet seriously impressive feats - and so the following are five of my favourite, culled from my new paperback and e-book, Siri For Dummies.

1. Reminders by location

It’s a breeze to ask Siri to remind you of something by time – such as “tomorrow at 10 am, remind me to call the dentist to book an appointment” – but you can also set reminders by location.

Related: Five ways your smartphone can save you money

For example, hold down the Home button on your iPhone 4S and ask Siri something like this: “Remind me to call mom when I leave here.” Whenever you leave wherever you are – such as your office, a coffee shop or shopping mall, Siri will remind you to call your mom. Your iPhone’s integrated GPS means it’s location-aware.

Alternatively, you can say “Remind me to take out the recycling when I get home.” Because Siri knows where your home is – or you might be asked to tell Siri this info once – you won’t be reminded of the chore until you pull into the driveway.

2. Siri can read your texts

You can dictate your text messages or emails, but  you can have your text messages read to you. Simply press and hold the Home button on the iPhone and say something like “Siri, read my texts.” Once Siri reads a message to you, you can say something like “Reply saying that’s an excellent idea, thanks” or “Tell her I’ll be there in 20 minutes.”

You can also ask Siri something like “Do I have any texts from “Mary”?

Related: How to extend your smartphone’s battery

3. Siri can handle Facebook, Twitter

While not designed to work with third-party apps like Facebook or Twitter, you can in fact use your voice to post status updates or tweets. It involves setting up these popular social networking services to work with text messaging.

Related: How to get paid apps for free

For Facebook, text the word “hello” to 32665 (which spells FBOOK on a traditional keypad). This works in Canada and the U.S. You’ll get an automated reply on your phone with a link to click to finish the setup. Finally, add the short code to your address book and rename it “Facebook.” Now you can tell Siri “Text Facebook,” followed by your status update.

It’s the same process for Twitter, as you’ll want to enable mobile updates by sending a short code to Twitter. Then you’ll add the short code to your Contacts, name it “Twitter” and then instruct Siri to send a message there (e.g. “Text Twitter, Did you hear there’s a Siri For Dummies book out now?”. To get going, text the word “start” to 21212 (Canada) or 40404 (U.S.). You’ll then receive a confirmation text right away.

4. It’s a verbal calculator, too

Siri includes support from Wolfram Alpha’s vast database of facts, definitions and even pop culture information (ask Siri who shot J.R. or Mr. Burns).

But you can also ask Siri to perform math problems for you, too. If you’re adding up cheques to deposit at the bank machine, for example, ask Siri something like “What’s $120.20 plus $345.12 plus $841.30 plus $572.21?” and within a moment you’ll hear the correct answer.

Related: This app finds deals near you

Alternatively, if you’re out with three friends at a restaurant and the bill comes to $300, you can ask Siri something like “What’s a 20 percent tip on $300, for four people?” and Siri will tell you everyone owes $15.

Naturally, Siri can also handle multiplication, subtraction, equations, fractions, and more.

5. Find your friends

If you’re not familiar with the free Find My Friends app, it taps into your iPhone’s GPS to provide your geographical location to people you choose to share this information with, such as a spouse, kids or friends.

Once you add consensual people to your list, you can see their whereabouts on a map, too, represented by a colored orb, and get the address they’re at, if desired.

You probably saw this one coming: You can use Siri to get the most out of the Find My Friends app. Press and hold the Home button and ask Siri “Where are my friends?” The app will open and you’ll see who’s around and how far they are from you. Now you can send someone a message like, “Let’s grab a coffee.”

You can also ask Siri something like “Is my husband at home?” or “Where’s John Smith?” or “Find my sister.” What are your favourite things to do with Siri? 


Feel free to share your experiences in the Comments section.

via http://www.speechtechnologygroup.com/speech-blog - Here are are a few commands you probably never heard off to get Siri do more for you..... Who knew phones could have a sense of humour? Siri, Apple’s voice-recognition software featured on the new iPhone 4S, has several witty comebacks for various questions. Handout/MCT One of the most talked-about ...

Friday, June 1, 2012

Ofcom: Telcos to make calls easier for disabled consumers - Recombu

How text to speech and speech recognition technology can modernize text relay services.

Ofcom is proposing an overhaul of UK telecoms to make things easier for those with hearing and/or speech impairments. Plans to upgrade text relay services, which converts speech to text and vice versa, were published last July. The proposal was that next-generation text relay services (NGTR) would:

“provide users with the ability to interrupt conversations, have two-way speech and the ability to benefit from the use of mainstream equipment. As with the current rules, our proposal was that all fixed and mobile communications providers should provide access for their customers to an NGTR service approved by Ofcom.”

Following on from July, Ofcom has published a list of performance indicators by which the effectiveness of NGTR calls would be measured.

From launch, NGTR would require voice to text transcriptions to be more than 98 per cent accurate, for a minimum of 40 words per minute to be received across calls with an average of 60 words per minute.

Ofcom: You cannot be Siri-ous?

Ofcom research has also found that existing text relay services are slow and ‘fail to flow naturally.’ It’s bad enough when Siri on the iPhone 4S doesn’t accurately compute our instructions - wilfully or otherwise.

Ofcom’s main aims are to improve the conversational flow and speeds and make NGTR available on mobiles, tablets, laptops and desktop PCs - so as well as over the phones, we’ll be seeing Skype-style NGTR services launching.

Ofcom has announced that it’s also consulting on video relay services, aimed at users of BSL (British Sign Language). Mobile operators blocking Skype may want to take note.

NGTR video calls will require “a broadband connection with data rates high enough to deliver high quality video.” Something which won’t be possible in the UK until the arrival of 4G or superfast broadband/FTTC in your area.

Ofcom concedes that video relays may not be a requisite for carriers, but is also weighing mandatory ‘implementation of [video relay calls] on a restricted basis,’ possibly in Wi-Fi-only areas like Apple’s Facetime.

Telcos and networks have until the 13th of July to respond to Ofcom’s proposals. You can read the full report here.

google.com by Thomas Newton
via http://www.speechtechnologygroup.com/speech-blog - How text to speech and speech recognition technology can modernize text relay services. Ofcom is proposing an overhaul of UK telecoms to make things easier for those with hearing and/or speech impairments. Plans to upgrade text relay services, which converts speech to text and vice versa, were publi ...