Since when did we start impulse to build and we as intelligent robots?
The summer of 1956, ten young scholars from different disciplines seminar held for two months, first proposed the concept of artificial intelligence. Since then, the "exploring the nature of intelligence," the romantic dream takes root in people's hearts. Make a like us "smart" robots also became the dream of the ultimate goal. According to the 60 people of the last century predicted that by 2020, the artificial intelligence of man can make his wisdom.
However, it is regrettable that, to date, our own how the brain works is still unclear, let alone make us like artificial intelligence. "Thorns, preventive and long" to describe artificial intelligence in this way, no way.
However, precisely because the face of each pit pit pit for the pursuit of artificial intelligence is more precious. Today, national intelligence and Daniel gather together to share their latest achievements in the field, also appears in a completely new direction of artificial intelligence.
Artificial intelligence, Alan Turing, the father of spirits
"Turing machine" out of it?
While Turing is a well-known homosexual, but "Turing base" is not as you think ... Turing was first proposed by Turing's computer model, which is based on the integration of abstract mathematics and logic, to complex calculations. , The "father of the computer" von Neumann design based on the first computer. Over the years, although the computer form has changed several times, however the kernel still does not change, that is based on mathematical calculations as the core.
Now, when you picked up the phone to shake a shake or looking at the beautiful woman self when you can't enjoy life, all thanks to Turing's old wet. However, Turing patterns to artificial intelligence planted a huge risk. Because our brains are not run according to the Turing machine model. Using Turing machines simulated human brain seems to make your Wang Xing who climbed trees, let your cats ball in mouth, filled with "the Minister qie can't do" feeling.
Now the computer and the human brain just how much of a difference does it make? Vice President Tan tieniu, academician of the Chinese Association for artificial intelligence one figure: the human brain contains neurons 10 11 orders of magnitude, contain neural node 10 14 orders of magnitude. And the laboratories in the most advanced transistors in the CPU can only reach 20 billion. This gap seems small, but on principle, every neuron in the brain is to collaborate, process information, respectively, while computers are smart to rely on a single, unified Computing Center. In terms of structure and brain function is a partition, while the computer does not function, but dedicated storage Center, operations center, as well as connecting bus between them.
In short, brain and computer organization is totally in both directions.
To see top this site pictures, this is a site General of New York Street, for existing of machine image recognition for, all of objects are is easy was defined, only left Xia angle this due to shutter had slow and virtual of of car most difficult was machine recognition, this in human mind in only need 0.1 seconds on can made of judge, in machine in the seems to is impossible completed of task. It has been using the largest computer to identify the image, still more than 1000 times of human time, and power up to 12 billion watts. By the way, the human brain only 20 Watts of power.
Need a new model of artificial intelligence?
But, traditional of artificial intelligence is not nothing, relies on big data and real networking, it can do many beyond brain of things, like by Germany industrial robot giant KAKU production of mechanical arm, on had and human to has one table tennis war, wants to see video can stamp here (table tennis game: Germany champion VS robot) but, this mechanical arm but distribution has frenzied of 36 only eyes (camera) Ah.
Back in 2011, Watson in IBM's artificial intelligence robots United States quiz show "Jeopardy" several times beyond the second achievement award.
However, rely solely on traditional artificial intelligence data and computation speed of progress is unlikely to last. IBM China Research Institute of the Soviet Union says:
IBM developed the most advanced chips are 7nm processes, and molecular-scale have reached 2nm, under existing technology, the transistor is unlikely to break the molecular scale. This means that the predicted chip-scale reduced by half every 18 months Moore's already close to its limit. Improve your computer's performance will also encounter bottlenecks.
Tianzi Jiang, Director of the Centre for the brain network group of automation Chinese Academy of Sciences told Lei Feng network, computer processing of information more effective, but the face of fuzzy thinking, the human brain is undoubtedly more effective, so the new direction of artificial intelligence needed from the human brain to find inspiration.
So, how do we build a computer to mimic the human brain?
In fact, this crazy idea not only now being proposed, as early as 20 years ago, scientists have proposed this method, however, when levels of hardware technology low explosive, so "brain chips" in recent years back people's vision.
Because Turing needs to continue to transfer data between the memory and arithmetic unit, and manufacturing "brain chips" must first get rid of such a structure. IBM SyNAPSE neural computing chip, eliminating the main line, CPU and memory data exchange, reducing a lot of power. At that time, the universal network of Lei Feng have been reported (IBM mimicking the human brain to create "thinking skills" chip) but don't think the chip is running in accordance with the way the brain works, and because humans are still don't know how the brain works. The chip runs only in respect of certain structure modelled on the brain, on the efficiency of ... We are also talking about something else.
IBM neural computing chips SyNAPSE
"Brain-machine interface" can be realized?
If I give you a cup of boiling water, you must let go the first time. If the information to the brain, and the brain after analysis, decision let go of instruction, and instruction on to the Palm, though only 0.2 seconds of the time, but you may have been injured. In fact, your arm is smart. These Smart also your nerves.
Artificial intelligence can not only learn from the human brain, also can learn from people's bodies.
This advanced concept has been used in artificial intelligence. Such as a computer or a cell phone linked to smart camera itself doing the screening of information, such as object when there is no movement within the vision, will remain dormant, as soon as motion is detected, then start recording and notification phone. A fridge can automatically detect it if there are eggs, whether food was the expiry date, which notifies the phone.
Horizon Robotics CEO Yu Kai from Lei Feng network
In the future of the world, people just don't realize the data, and produce data to all sensors and cameras. All smart times there will be tens of billions of Internet users, with more than 6 billion people, the rest of the robots. Each has its own separately Smart network hardware capabilities.
However, these were not enough burst, Hu Dewen to tell of Lei feng, a professor at the University of science and technology for national defense: "if artificial intelligence can in some ways really simulates the operation of, the next step of artificial intelligence direction must be brain-Internet. " Victorias Secret iPad Mini Case
Enter human visual and auditory information, according to mathematical models of brain-wave formation can be machine readable, so as to achieve the purpose of brain waves to control a computer, is the basic logic of brain-Internet. Although scientists can signal resolve the subjects mind trying to spell the letters, but nothing more. Because the human brain waves are very weak, abstracted visual signal is a very difficult thing. This reached also the cutting edge of artificial intelligence research. It is said that Chinese scientists are trying to use radio control cars, have successfully knocked several police. (Small make up just want to ask, can you find driving shoes emitting brain waves ... ）
Speaking of which, or reference a Turing-Word
We can only see the future not far away, but within such a short distance, we still have some work to do.
We learn from birds, and invented the airplane, learning from the bat, the invention of radar, and now finally to their own study. Don't know if you have found that no matter how expensive the computers before you are far from your own precision. In fact, most expensive computers that you own, have such a high body, will continue to waste your life? So, get up and move bricks! Victorias Secret iPad Mini
Tips
Micro sweep sweep, author tips bar ~
1530 votes
Galaxy S6 edge
S6 Edge above Exynos 7420 processor Samsung aligned currently dominate the market for high-end mobile processors of Qualcomm's crotch hard feet, 14nm FinFET technology, as well as the Samsung 6 generations perfecting the design of the size of Samsung Snapdragon 810 processor 7420 phase went on sale now has an absolute advantage.
View details of the voting >>
No comments:
Post a Comment