VTUbuddy !

Google Brain has learnt recognising Cat Images.

catThe next time a cat “mews” do not shoo it away. On the contrary try to reflect the great upheaval it has created in the lab of the greatest search engine of the world GOOGLE. Yes,Google with its heuristic purpose of developing AI has created one of the largest neural network for machine learning by connecting 16000 computer processors which

they turned loose on the internet to learn on their own. The purpose was to make understand images(in this research it’s the cat) by machine automatically. The effect of the creation was some activities what human brains do-that is look for cats. This is what Google’s brain did with 10 million digital images in Internet,that is learn by themselves. 

The Google brain assembled a dreamlike digital image of a cat by employing a hierarchy of memory locations to successively cull out general features after being exposed to millions of images. The scientists said, however, that it appeared they had developed a cybernetic cousin to what takes place in the brain's visual cortex. Currently modern mechanistic vision technology is being used for the learning process by labeling certain features. But ,however, in the Google research no identifying features were fed to the machine. The idea was that to scatter innumerable data to the algorithm and let the data take on the driver’s seat and have the software automatically learn from the data. The machine was never told that “it is a cat”,but it meticulously invented the concept of cat. 

Humans understand familiarity through repetitions. In brain the specialized neurons are triggered off when familiar images are exposed repeatedly. A fitting analogy was being established in the Google search by grafting numerical parameters with synapses where exchange of stimulus takes place. The memory region of the computer was filled up with parallel images of cat and human body parts. The scientists were cautious about drawing parallels of human biological system to that of software system. 

General Intelligence is still the long term proposition of under the preview of AI. AI includes highly technical and specialized issues having various branches. The branches are nurtured under specific institutions or under specific researchers. The branches deal in finding some specific solution to one of the many possible approaches of aproblem. Presently AI focuses on the reasoning , planning , learning , communication ,perceptions etc. AI has received many accolades and had been a subject of optimism but also suffered major setbacks as genres. Today however it is an essential part of the technology industry giving major boost for most difficult problems in computer science. 

Google X Lab has remained a very secretive part of Google of which most of its employees are not aware of. Talented brains are being acquired and provided necessary emoluments to carry on path breaking researches in Google’s X Lab. One of the main mission of this lab is to find out the proper algorithm to initiate AI in full force. The long term goal for AI has been in Google’s plan for quite a substantial amount of time. Google’s cofounder Larry Page has often voiced out Google’s dream run for AI. Google had always wanted to acquire the biggest global lab for AI. Also the affinity for AI has remained in the major job contents offered by Google. However the inception of the idea and its execution through Google X lab has remained an enigma. The  researchers there are not even given the crystal idea about the purpose and the works are not done on regular basis. They meet casually at intervals and the meetings are clandestine and secretive.The lab was mentioned publicly in 2011. The investment done in the lab is also nominal compared to Google’s core business. The lab takes over special projects which deal with metaphysical ideas and turn them into products which affect and impact the society in the long run and aids for its progress. 

 Some of the important projects of the Google X lab can be of special mention. The products have not been yet open to public and product testing is in view. The major projects are 

1)Project Glass - a research program to build an augmented reality head-mounted display. The first demo resembles a pair of normal eyeglasses where the lens is replaced by a head-up display. 

2) Driverless Car - this is a unique project with a self driven motorized vehicle which has been put to test but not yet been marketed. 

3)Google Assistant / Majel - a voice recognition and comprehension system to be used as a virtual assistant in Android phones. This is Google's equivalent to Apple's Siri. The project was known as "Majel" before March 2012. After that it became "Google Assistant” 

There have been other researches also. Larry Page and Sergey Brin came up with a list of many ideas, including: Robotic avatars, teleprocessing, for telecommuters and meetings,Web-connected light bulb communicates with Android - was said to be unveiled by end of 2011(however its not through yet)  and Space elevator which had been a long-held dream of the Google founders. 

However along with these researches,a small group of scientists at Google X lab dreamt of a small group of researchers began working several years ago on a simulation of the human brain. Although some of the computer science ideas that the researchers are using are not new, the sheer scale of the software simulations is leading to learning systems that were not previously possible. The neural network taught itself to recognize cats, which is actually no frivolous activity. This week the researchers will present the results of their work at a conference in Edinburgh, Scotland. The Google scientists and programmers will note that while it is hardly news that the Internet is full of cat videos, the simulation nevertheless surprised them. It performed far better than any previous effort by roughly doubling its accuracy in recognizing objects in a challenging list of 20,000 distinct items. 

However Google scientists are not the only one in the race. Last year a group of Microsoft’s scientists showed that similar artificial network can be used to voice recognition technique. Despite their success, the Google researchers remained cautious about whether they had hit upon the correct combination of machines that can teach themselves. Although the breakthrough is simply a great achievement,it needs to be seen whether the process can be made bigger with getting the exactly correct algorithms. And do not forget the cats!


You are here: Home Articles Google Brain has learnt recognising Cat Images.
Follow Me on Pinterest
Latest Pins: