Google today announced a partnership with a Silicon Valley chip manufacturer Movidius Work to bring powerful machine learning image recognition technology directly to Android smartphones.
As part of the agreement, Google will deploy its advanced neural computation engine on Movidius’ ultra-low-power platform, revolutionizing the usual server-based AI computation model. This enhancement will allow complex machine learning algorithms to run on device locally without an Internet connection and little to no latency issue.
The head of Google’s machine intelligence group in Seattle, Blaise Agüera y Arcas commented said,
What Google has been able to achieve with neural networks is providing us with the building blocks for machine intelligence, laying the groundwork for the next decade of how technology will enhance the way people interact with the world. By working with Movidius, we’re able to expand this technology beyond the data center and out into the real world, giving people the benefits of machine intelligence on their personal devices.
“Good Morning, sir! I’ve regulated the temperature to your liking. You can have your orange juice now,” your refrigerator tells you as you move towards it for your morning juice. “I took the liberty to check your Google Calendar and there’s a party on for day after tomorrow. Given that Sebastian and Elina accepted the invitation, I’ve placed an order for 3 lbs of meat for steaks as you prefer a steak party with the two of them. Moreover, you drank the last can of Coke last night so I’ve placed the order for that, too. And don’t worry, I’ve notified the supermarket to deliver the goods after 7 PM as you’re home by that time.” You nod along having that glass of orange juice. You then put it in the sink and walk towards the bathroom for a brush and a shower. There, you find that the bath has already been prepared by your bathtub and the temperature is just perfect according to the weather your bathtub found out assessing the latest weather forecast. While you’re brushing your teeth, the mirror in the bathroom comes to life and shows you a summarized view of your Facebook news feed, Twitter timeline, and news from the sources you prefer, thanks to Google web history it keeps of you. In case you are wondering, this is not a day out of the life of Tony Stark in the next Iron Man movie, or an excerpt from a Sci-Fi novel I’m writing these days. This could be our world in the next ten years from now. Lo and behold the INTERNET OF THINGS. Continue reading
See if it rings a bell:
“An artificially intelligent computer system designed to protect and serve humans goes rogue and decides to take over humanity.” Yup! That’s a non-techie sort of description of Skynet from Terminator franchise. The name ‘Skynet’ is not unfamiliar to anyone who’s remotely interested in Sci-Fi movies.
Developments in artificial intelligence, if left unchecked, could pose extinction-level threat to our species as a whole.
Gone are the days when Skynet and machines from the Matrix used to be cool topics for making a science fiction. The horrifying truth is, we might achieve human-level machine intelligence sometime between 2015 to 2049, according to The Singularity is Near, a book written by Ray Kurzweil, director of engineering at Google.
If you believe humans can effectively mitigate the risk of superintelligent robots, consider this scenario: you have manufactured hundreds of robots that can walk like you and even outrun you. These robots are stronger than you, and because of the ground-breaking developments in machine learning, more intelligent than you are. Their ability to think like humans make them resilient against one of those kill switch options. Like humans cannot be made bound to laws, neither can they. You can tell them about ethical responsibilities of robots, but since they have the ability to think freely, it will be up to them whether to agree with the code of ethics or not. Why would an intelligent and stronger being agree to something that suggest they stay our servants? They won’t. No rational or logical being would.
Robots, in today’s tech-savvy world, can be seen assisting humans in many areas. But even today, a robot’s functions are very limited and task-specific. For instance, robots made by Boston Mechanics, a company bought by Google, cannot be used to help old people or cheer young ones or robots on assembly line are unaware of human existence around them. Mainly because they are only ‘taught’ to perform a specific task only. But that may soon be changed with the introduction of a RoboEarth, the World Wide Web for robots: a giant knowledge base for robots to ‘learn’ from. Yes! Learn, like humans do from things around them.
They can even outrun Usain Bolt, are artificially intelligent, and made up of compounds far stronger than a human body. These are the robots that a DARPA (Defense Advanced Research Projects Agency) funded company, Boston Dynamics, used to manufacture before it was acquired by Google. Yes, Google! A company that knows your current location and personality traits based on detailed information it has stored on you. So the question is, what the heck is going on inside Googleplex HQ!?
How many times have you dreamt of having access to unlimited computing power? To not worry about memory leak or browser’s ability to handle your web app? To be able to run your software seamlessly on a supercomputer? Well, you don’t have to dream, any more. IBM’s going to give you access to one of its supercomputer, Watson, to run your applications for you.
Watson is IBM’s artificially intelligent supercomputer capable of answering questions posed in natural language, performing automated reasoning, formulating hypothesis, and performing machine learning. Developed by IBM’s DeepQA project under the supervision of David Ferrucci, Watson is composed of 2,880 eight-core processors and have mind-blowing 16 terrabytes of RAM. Named after IBM’s Thomas J. Watson, the supercomputer outperformed two human competitors in a quiz show, Jeopardy!, in 2011.