The human brain, fascinating, exciting and full of possibilities, the notion to create, form an opinion and challenge the environment which we live in, is truly exceptional. We now might be able to find answers as to how powerful the human brain is after a project which is designed to compare a supercomputer with that of a brain.
An Artificial Intelligence project which has been devised by two PhD students from the University of California Berkeley and Carnegie Mellon University, will be the first of its kind to compare the human brain with the world’s best supercomputer. The AI Impacts project aims to determine how fast the human brain sends signals in its internal network compared to that of a supercomputer.
The scholars compared the power of our brains with that of IBM’s Sequola supercomputer which is in the top 3 of the most powerful supercomputers. “Sequola has a TEPS (Traversed Edges per Second) benchmark of 2.3 x 1013 TEPS”. The estimates suggest the “AI Impacts are that the human brain should be at least as powerful as Sequoia in the lower limits and for the upper estimates, therefore the human brain could surpass the IBM Sequoia speed by 30 times at 6.4 x 1014 TEPS”.
Which is both a lot to take in but also equally and potentially incredible, evolution has formed an instrument which is quite amazing, and it begs the question, what else will we find as research and tech advances with the aim of exploring us. It is also interesting to note if the wiring of for example a genius brain, think Stephen Hawking, is different to that of an average mind or the best sportsman evolved differently with more advanced genes, or if are we all capable. If we spent enough time learning a skill to be able to adapt to anything? Its compelling none the less.
Thank you aiimpacts for providing us with this information.
Image courtesy of fossbytes
tl;dr: in a few years we’ll have supercomputers more powerful than the human brain.
According to the Graph 500 site (http://www.graph500.org/results_jul_2015) the current fastest computer has 38621.4 GTEPS (3.8 x 1013 TEPS).
Notice the x 1014 in the human brain test?
Let alone the 6.4.
I find that really hard to achieve in a few years only, but I think it will happen eventually.
Notice the x 1014 in the human brain test?
Let alone the 6.4.
I find that really hard to achieve in a few years only, but I think it will happen eventually.
Notice the x 1014 in the human brain test?
Let alone the 6.4.
I find that really hard to achieve in a few years only, but I think it will happen eventually.
Notice the x 1014 in the human brain test?
Let alone the 6.4.
I find that really hard to achieve in a few years only, but I think it will happen eventually.
Notice the x 1014 in the human brain test?
Let alone the 6.4.
I find that really hard to achieve in a few years only, but I think it will happen eventually.
The ratio is very simple. If you assume that human brains indeed have 64 x 10^14 TEPS and that it’s a measure of brain power (these are the main articles for dispute here), then a human brain has about 28 times the power of the Sequoia and 17 times the power of the K computer.
A year ago the fastest computer had 18,000 GTEPS, 2 years ago 15,000, 3 years ago 3541, 4 years ago 18 GTEPS. It shouldn’t take long for the fastest computer to be 17 times faster than it is today.
(Well, that may not be accurate. It represents the computers tested with the benchmark. It may be that the testing wasn’t done on the most powerful computers 3 or 4 years ago. Still, even assuming a growth of only 2x every 2 years, within 8 years the most powerful computer will be about as fast at this benchmark as the human brain.)
I have been checking out some of your posts and i can state awesome and nice stuff. I am definitely bookmarking your blog.
IPL 2017 Team Player List