And
it's not a criticism of Google, according to researchers in the field
of artificial intelligence Jann LeKun and Josh Tenenbaum. Scientists
were held during the New York World Science Festival after the premiere
of the movie "Creator: Alan Turing and the future of thinking
machines," an extraordinary art-house films about the perception of
artificial intelligence by Alan Turing in the 40-50s of last century.
World Encyclopedia, known to us under the name of Google, is commendable in many respects, in particular, the amount of information that it can absorb and give a response to almost any request. Yet this is only an electronic computer learning method of "memorizing" when there is no understanding of the full notes LeKun, a professor of computer science and neuroscience at New York University.
With respect to computer skills, even the most powerful computers of the world only close to the capabilities of the insect shows a scientist. "I would be happy to catch during the life of a computer on the uptake corresponding to the rat," - he said.
Seemingly amazing things that are available on Google, such as near-instant provision of information on the direction of motion, using only an elementary form of artificial intelligence called "simple plan". "It's easy - says Tenenbaum, a professor of computer science at the Massachusetts Institute of Technology. - It is not even called artificial intelligence. It's just Google ».
As the scientists explain, the real reason - it is not just remembering, but also to use this information to understand the situations that you have not encountered before. An example of what can only be seen by the film. "You're watching a movie and see the images that you did not know before. You may not know anything about the life of a hero - says Tenenbaum. - Here is the context and informed exchange of information, understanding the inner world of another person just by what he says, and you are responsible - this is the essence of the wisdom of man. "
Two professors are exploring other areas. For example, LeKun experimenting with moving the robot, which tries to detect surrounding objects. The scientist showed a video that shows what the robot sees, as it classifies objects such as people, trees and roads. Typically, the machine properlyperforms tasks, but often refers to a robot tree people, water stains, and the building of a lamppost.
Available to explain what they are dealing with the researchers in the field of artificial intelligence, LeKun described the magnitude of the human brain based on the most recent, but very rough calculation: 100 billion neurons are generated by a 000-10 000 connections with other neurons, and use these compounds from 100 to 1000 times per second (rather inflated data). As a result, we may quintillion - 1,000,000,000,000,000,000 - transactions that occur in the human mind every second.
"But at the same power of supercomputers is increasing exponentially - LeKun said, noting that they have achieved the above capabilities within 30-100 years. - Even 10-20 years - and will find and smartphones. That's because your phone will be smarter than you. "
World Encyclopedia, known to us under the name of Google, is commendable in many respects, in particular, the amount of information that it can absorb and give a response to almost any request. Yet this is only an electronic computer learning method of "memorizing" when there is no understanding of the full notes LeKun, a professor of computer science and neuroscience at New York University.
With respect to computer skills, even the most powerful computers of the world only close to the capabilities of the insect shows a scientist. "I would be happy to catch during the life of a computer on the uptake corresponding to the rat," - he said.
Seemingly amazing things that are available on Google, such as near-instant provision of information on the direction of motion, using only an elementary form of artificial intelligence called "simple plan". "It's easy - says Tenenbaum, a professor of computer science at the Massachusetts Institute of Technology. - It is not even called artificial intelligence. It's just Google ».
As the scientists explain, the real reason - it is not just remembering, but also to use this information to understand the situations that you have not encountered before. An example of what can only be seen by the film. "You're watching a movie and see the images that you did not know before. You may not know anything about the life of a hero - says Tenenbaum. - Here is the context and informed exchange of information, understanding the inner world of another person just by what he says, and you are responsible - this is the essence of the wisdom of man. "
Two professors are exploring other areas. For example, LeKun experimenting with moving the robot, which tries to detect surrounding objects. The scientist showed a video that shows what the robot sees, as it classifies objects such as people, trees and roads. Typically, the machine properlyperforms tasks, but often refers to a robot tree people, water stains, and the building of a lamppost.
Available to explain what they are dealing with the researchers in the field of artificial intelligence, LeKun described the magnitude of the human brain based on the most recent, but very rough calculation: 100 billion neurons are generated by a 000-10 000 connections with other neurons, and use these compounds from 100 to 1000 times per second (rather inflated data). As a result, we may quintillion - 1,000,000,000,000,000,000 - transactions that occur in the human mind every second.
"But at the same power of supercomputers is increasing exponentially - LeKun said, noting that they have achieved the above capabilities within 30-100 years. - Even 10-20 years - and will find and smartphones. That's because your phone will be smarter than you. "
No comments:
Post a Comment