Understanding Requires Intelligence

A new survey from Intelligent Voice says that 46% of people think that Apple oversold Siri’s voice capabilities. Forgetting for the moment that this actually only represents the views of 69 Americans, a minutely tiny number compared to the estimated 250 million downloads – and hence users of Siri – that already have iOS 7, I would argue that, if anything, Apple has underplayed the importance of voice.

As wearable technology takes off and consumer devices become more interactive, voice will be a key input method. And not just for use as simple command and control or keywords, but as an interface that allows complex sentences to be used as people truly start talking naturally to their devices.

To deliver the expected experience it will take much more than the basic understanding that ASR technology gives.

That is because ASR just converts spoken word to text. It can’t intelligently understand the context of a conversation, predict a user’s intention or intuitively interpret what the user really meant. To do that, you need NLI! NLI enables this and more, allowing the machine to understand the user easily and generate the correct response accordingly, such as providing an answer to a query or carrying out an action.

I don’t recall, as the survey suggests, that Apple promised “Star Trek” communications. What the survey is missing is that this is a journey and there’s no doubt that Siri – and other NLI based technologies such as Google Now and Artificial Solutions’ natural language app, are improving in leaps and bounds. It’s interesting for example to look at the feedback to the recent smart watch releases that so far seem to have been met with a fair degree of disappointment in that the features they offer are pretty limited. In many ways, I view that as a positive since it shows that the demand for “things to do more” is well and truly there.

Look at this way, without voice as a primary input medium, much of the next generation of technology such as smart-watches, Google Glass and its equivalents and a whole tranche of new wearable technology is going to be pretty pointless. It’s bad enough trying to type out emails on a mobile, let alone interact with an even smaller screen!

So, is speech enablement perfect right now? No, but it’s already come a long way and technology such as NLI makes up for the deficits in ASR as well as providing the necessary artificial intelligence to deliver true natural language’s potential. Is it worth perusing? You bet it is.

Communication was probably the most important evolution for humans; imagine what we can achieve by being able to talk to machines.

 

Leave a Reply

Your email address will not be published.

top