IVR Deconstructed

High functioning natural language processing is hands down one of the most sought after features of interactive voice response systems.  Natural language processing (NLP) is a field that combines computer science and linguistics and studies interactions between human language and computer processing.

As the above title suggests, the nuances and slight variations included in every human language make this one of the most difficult and confusing fields in all of science.  How would one teach an inanimate object that doesn’t have consciousness or inherent lingual abilities to distinguish between two words that have very different meanings but sound phonetically the same?

NLP is a sub-branch of artificial intelligence, and is primarily focused on developing technology that allows machines to interpret verbal communication that enables humans to speak in their natural language, dialect, and accent.

So where did this idea of natural language processing originate?  Presumably, computers have only been functional on a consumer level for several decades.  It actually dates back much further than most would assume, and can be traced to patents applications in the 1930s for a translating machine.  The machines would in theory be bilingual and able to translate between languages with varying levels of complexity (including dealing with grammatical roles between languages).

Actual software capable of performing NLP tasks was not released until 1954 during an experiment at Georgetown University and IBM.  Researchers were able to program an application to fully and automatically translate more than sixty Russian sentences into English.

There was a ten year gap between the development 1954 and 1964 (when an application was released that could solve algebraic word problems), put post-1964 there was a steady stream of NLP applications released.  Some of the more interesting ones included chatterbots able to simulate natural human language and generate prose at random and a question answering system that was so advanced that it was able to win Jeopardy when pitted against some of the best human contestants in the world (most will recognize the device as Watson).

When NLP programs are integrated into software applications and IVR systems, the results are lauded as the type of programs that science fiction would imagine.  The actual technology behind NLP applications is extremely complex and requires extensive research and testing to implement.

NLP software and applications have been developed that perform a variety of language interpretation tasks.  Various tasks include translating text from one human language to another, identifying the discourse structure of text, converting information stored in databases to readable language, offering answers to human-language questions, and even identifying the sentiment of spoken or written text based on extracting subjective information.

Coming up: The challenges and complexities of programming a machine to understand a person.

Freedom author Jonathan Franzen thinks ebooks aren’t permanent enough, that the digital form doesn’t have the permanence of print. He’s talking about the feel of a book printed permanently onto the pages.

It may be a surprise (I do work for a company selling technology—interactive voice response systems—after all), I agree with him. Unfortunately, I think ebooks are the books with the real longevity. And I don’t think it’s a question of which we like better.

“Maybe nobody will care about printed books 50 years from now, but I do,” Franzen told an audience at the Hay festival in Cartagena, Colombia, according to the Guardian. “When I read a book, I’m handling a specific object in a specific time and place.”

I get that. There’s a feel of permanency you just don’t get from a screen that three seconds before was showing you a YouTube video of some kid picking his nose or whatever.

“Someone worked really hard to make the language just right, just the way they wanted it,” Franzen said. “They were so sure of it that they printed it in ink, on paper. A screen always feels like we could delete that, change that, move it around.”

Preaching to the choir, buddy. One of my favorite books is an old hardback copy of War and Peace that’s been in my family a few generations. The cover is torn in places and the pages are yellowed with age. It’s about 70 years old, I think—my grandmother’s originally.

Reading it, I can feel the passage of time in it. The age enhances the experience of the novel, which was written over 150 years ago. It feels different than the paperback copies of War and Peace I pick up at the bookstore. I don’t sense any age in a new paperback version, let alone on a digital screen.

But in the end it’s a question of morals—or it will be. At some point we’ll have depleted our forests enough that paper books will be unthinkable. Our morals as a society will shift so that killing trees to make books will be frowned upon and then outlawed.

So as much as I agree with Franzen on his point about ebooks, I don’t think we’ll have much choice in the matter. Printed books will disappear. Ebooks will be the only books we know.

We have a conundrum on our hands. With the rise of social media, anyone with an Internet connection can spread their views all over the globe. Which poses a problem, for free-speech nations as well as suppressed-speech nations.

Today, Twitter announced that it will censor tweets if governments ask for it. It’s kind of a bandaid approach to a much bigger issue, which is that social media has opened up the whole world. Can we control it? Should we?

Twitter’s removal of tweets upon request by governments is targeted towards free-speech nations, not oppressive ones. It’s intended for nations like France and Germany, who are constantly working to suppress anti-Semitic or pro-Nazi sentiment.

Twitter doesn’t operate in suppressive nations because of the fundamental incompatibility inherent there. It just doesn’t work. By its very nature, Twitter requires freedom of speech.

“We hold freedom of expression in high esteem and work hard not to remove tweets,” Twitter spokeswoman Jodi Olson told CNN. “And just to be clear, this is not a change in philosophy, and there are still countries to which we will not go.”

According to the Vancouver Sun, Twitter can’t get into China, for example. China’s censorship starts with what’s referred to as the “Great Firewall of China.” Global Internet service enters through eight government-filtered gateways. Filters include long lists of websites and keywords that Chinese Internet users can’t access—they receive a “page cannot be found” error message.

That’s one way to do it. Another way is the Bavarian approach to Mein Kampf, Adolf Hitler’s manifesto from 1923. Basically, the Bavarian state government, which owns the rights to the work until 2015, has blocked publication of the controversial work over the years. There’s actually no German law on the books blocking it, Bavaria has just refused to allow it.

Wednesday, a German court blocked publication of excerpts from Mein Kampf in a British academic work, according to CNN. It’s effectively only a three-year ban because the rights expire in 2015, but it’s a statement nonetheless. I mean, anyone can read Mein Kampf on the Internet (in most countries, of course).

But it’s not like censorship is new. Even free governments have influenced or controlled the media at various points in history, particularly during wartime.

A May 10, 1945, article in The Miami News ran a story about a wartime correspondent up in arms about censorship of his Reims surrender story, for example. Correspondent Edward Kennedy called it “purely political censorship,” saying that military security wasn’t an issue.

In 2011, we saw the power of social media in the Arab Spring. It became a primary weapon for revolution. It’s an issue that’s starting to come up more and more.

Every government asks Twitter or Facebook to censor something from time to time. Those companies actually keep records of the requests, which anyone can look up on the Internet (again, in most countries).

I guess the question is: Is it right to censor views, even if those views are abhorrent to most of us? I don’t know. But I think we’re just beginning to see this emerge as an issue, and it’ll grow in importance as social media continues its rise.

Three—

A couple years ago, Google got busted in a federal sting operation that involved a convicted con artist, advertising for illegal online pharmacies and a $500 million settlement. Crikey, it sounds like a movie script.

According to the San Francisco Chronicle, the feds used a “convicted fraudster,” in prison at the time, to pose as a businessman running illegal online pharmacies.

At the request of the feds, grifter David Whitaker set up fake businesses pretending to sell first steroids and human growth hormones and then “prescription narcotics, Prozac and Valium without prescriptions.”

Whitaker recently told the Wall Street Journal that Google’s ad executives didn’t go for it initially, but then they worked with him to try to circumvent the laws governing this type of activity. Oooooh, that’s totally Dark Side.

From the San Francisco Chronicle:

Most damning of all, Whitaker told the Journal that he made phone calls in which he walked through the illegal parts of his fake Web sites with Google execs, and once told them that one of his clients wanted to become the biggest steroid dealer in the U.S. The calls were recorded.

All this happened two years ago but is news again after Whitaker talked to the Wall Street Journal. The timing on it couldn’t be worse for Google, what with the other Dark Side things on the company’s dirty laundry list in the news right now.

People are jumping on the hypocrisy of it all considering Google’s slogan, which now just seems to be a tongue-in-cheek inside joke—either that or the company just doesn’t operate under the same guiding principles it used to. No matter what, it’s definitely Dark Side. Yoda would never do that.

But there’s still more…

Four—

Google is being roundly criticized left and right all over the newswires and tech blogosphere for changing their security policies and giving their users a take-it-or-leave-it choice.

Basically, Google users have to accept sharing of their information between many of Google’s various products (Gmail, Google+, et cetera). It’s either that or users close their accounts—they can’t choose what information they want to share.

You know who else gives take-it-or-leave-it choices? The Emperor, that’s who…

“If you will not be turned, you…will…be…destroyed.” (In that waaaay creepy voice.)

Don’t do it, Luke. The Empire might have better pay and bennies, but you’re gonna hate the work. Same to you, Google.

The analog way for families to keep track of soccer practice, ballet and flute lessons is a calendar tacked up on a wall in the kitchen. But, as we all know, we don’t live in an analog world anymore.

If families were like companies, they could keep track of things with iCalendar or Outlook, but that would require laptops or smartphones for everyone in the family, along with some tech savvy.

A study by Iowa State University researcher Mark Monroe proposes something much simpler—using a voice interface system to do the job.

“[Feature] cell phones have enabled communications between family members but don’t provide access to systems such as email and the family calendar,” Monroe writes in his report, Remote Voice Interface for Home Communication Tasks.

Let’s face it, what dad or mom wants to buy $600-$800 smartphones for every member of the family? A much less expensive and simpler approach could be to set up a voice interface system.

“The proposed system is a home information system focused on improving family communication and control of household tasks. A core set of communication functions includes voicemail, email and calendar.”

With the advances in recent years in voice-interface technology (particularly the advent of VoiceXML IVR and improvements in speech recognition software), such a system is possible.

“In order to make these functions as available as possible, they must be accessible in various forms, via various means.”

What Monroe proposes is a centralized system that resides in the family home but is accessible from a variety of devices. Family members could access the calendar or emails on the family home computer, via a PDA, by calling into the system for audio messages or even with a “portable screen in the kitchen.”

It’s an interesting concept—kind of like the systems used in the corporate world, just scaled down for the family.