« How Digital Technology Can Replicate the In-Store Experience | Main | How the Glass Could Revolutionize the Business of Insurance »

May 28, 2014

Rooting for Computers That Think Like The Brain

Posted by Dr. Srinivas Padmanabhuni (View Profile | View All Posts) at 12:41 PM

Google's Artificial Intelligence Plans [Source: https://www.youtube.com/watch?v=As4fXoI8WpI]

Science fiction movies are filled with fascinating references to mankind's distrust of artificial intelligence. Remember HAL from 2001: A Space Odyssey? The computer quietly observes human behavior until it reveals to the beleaguered crew that it is indeed in charge of the spaceship. The Matrix was another such tale. It took the protagonist a lot of cajoling to convince him that he lived in a computer-generated reality. That computer, like HAL, didn't want to cede control of the situation.

We're now at the point where we don't have to go to the movies to gauge our distrust of artificial intelligence. A case in point is the "captcha," the often difficult-to-decipher list of letters and numbers that is supposed to discourage spammers and automated entities from getting access to Web sites. A group of engineers invented the captcha more than a decade ago as a way to distinguish between humans and computers on the Web. But what might not come as a surprise is that an artificial intelligence company claims to have figured out a way for its computer to crack most captchas. The computers once again appear to be beating us at our own game!

Whichever side you're rooting for (and you might very well be partial to computer brains if you are an engineer), one thing is certain: the fact that artificial intelligence is advancing to the point where it can think like a human brain is pretty amazing. It's part of a technology movement that aims to make computers see not unlike our human eyes do. And for that to take place, computers still need to bridge a large gap between numerical processing and sensory perception.

Until now, a computer was only as good as its programmer. As long as it confronts problems from a series of computations based entirely on 0s and 1s, there's very little a modern computer can't do ... and in record time. But when you start asking even the most advanced machines to classify a spoken statement as either serious or a joke, you start to test its limits. I was reading a fascinating article that referred to new software from Google that can distinguish animal faces from human faces when presented with various video clips. Apparently it takes 16,000 processors to power this sort of software.

Clearly we're on the path to producing computers that can "think" more in line with humans. That we're on our way to building machines with neural synapses and cognitive functions is inspiring. But what are the commercial possibilities? It's neat that my laptop someday might be able to "listen" to a piece of music and identify its composer. This technology truly goes a long way, however, when it comes to the driverless car. That prospect would become a lot more realistic if the computer platform powering those cars had cognitive skills - the ability to distinguish a small child running into the street from a plastic bag that's blown into the path of the car by a gust of wind. The former is a situation in which the brakes must be applied; the latter is just an empty bag blowing in the wind. It's easy for us to recognize the difference, but difficult for even the most sophisticated of computer platforms to judge time and again on the open road.

That's why I'm rooting for the machines. I think enterprises of all stripes, from automobile manufacturers to retailers, will benefit from computing platforms that can see the world as their human customers do. Whether it can drive a car or drive in-store sales when a shopper is deciding what shirt or shoes to buy, a computing platform more closely resembling the human brain will aid companies in their quest to better connect with digital consumers. cognitive computational platforms are now being evaluated for applications in domains like healthcare and insurance.

It's amazing that computers can now go on and win game shows and successfully challenge chess grand masters. Yet those same engineering marvels might have trouble recognizing which of three modernist paintings is a Picasso. The point we're at with computational intelligence is not unlike where we were with Big Data only a few years ago. A lot of that data existed; we simply hadn't yet figured out an efficient way to process it so that we could crack into the minds of our consumers.

In specific case of services innovation , cognitive computing has the potential to disrupt human -centered processes and information technology services with the emergence of robotic automation.

Computers that see, smell, and hear are on their way, and the commercial applications of such machines will be limitless. So much for the captcha...

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Search InfyTalk

+1 and Like InfyTalk

Subscribe to InfyTalk feed

InfyTalk VBlogs: Watch Now

Infosys on Twitter