-
Add Bookmark | Recommend this book | Back to the book page | My bookshelf | Mobile Reading
Free Web Novel,Novel online - All in oicq.net -> Sci-fi -> Lost in the stars Volume 4: Walking in the Solar System Chapter 52: New Intelligent Brain
Previous page Return to Catalog Next page
- Li Zhuang is doing his father's duty very leisurely, but his disciples are still working overtime today. The core technologies of the Artificial Intelligence Research Institute were built by teachers. The core technologies of Bai and Rai come from "robots." Therefore, there are many students from both countries studying in Weizhou, many of them outstanding, which puts them under great pressure. If the inner sect disciples lose to the outer sect disciples, these guys who think highly of themselves will not be able to see anyone.
The L-level is estimated to be the most produced six-level brain on the planet. It and the V-level, which is one level higher, are very special because of Li Zhuang's preference, and the emotional simulation circuit is complex. Everything has pros and cons, so L-level and V-level have become a heavy burden for Li Zhuang. He spends two hours every day communicating with the smart home, pondering the variables and checking the operating status. Dr. Luke Macaulay, the boss of hsai, announced his retirement on New Year's Day next year last month. He is only five years older than Li Zhuang and is still in his prime period as a scientist, but he can no longer tolerate the heavy after-sales service.
Wang Chengmin was surprised that the teacher had such weird ideas, because Li Zhuang did not regard the brain as a tool, but as his own children. This design scheme is obviously not suitable for mass production. Guster's idea coincides with his senior brother. The new smart brain project team led by the two began to design a maintenance-free dedicated six-level smart brain with better performance six years ago.
With the joining of junior brothers Jing Ran and Terry Weibel, the development speed of this new intelligent brain, codenamed nl level, suddenly accelerated. Both Jingran and Weibel's qualifications are not inferior to Guster's, and the younger Zhen Yun's qualifications are slightly higher, which means that his IQ may be higher than Li Zhuang's, but he is too cheerful. If he can't change in the future, I'm afraid he won't be able to bear it. Boring long hours of research. Qian Lin is proof of this. He has a great life at the Intelligent Life Research Institute, but this field is not the mainstream of AI scientific research.
The Artificial Intelligence Research Institute has done in-depth research on the transition between quantitative and qualitative changes in artificial neural networks. The intelligence level of the new brain is 6.0. It also uses one million neurons. The system architecture, algorithm and data structure have been improved, eliminating L-level complex emotion simulation software and circuits. It has almost no self-upgrade potential, but can ride on the threshold after standard training mode. It is a pure tool, requiring no communication and no need to consider variables. Ordinary people can communicate freely according to the standard mode and there will be no risk of privacy leakage.
After three years of R&D and three years of improvement and testing, the NL-level was officially completed in April. It was originally planned to be named the WG-level. Wang Chengmin considered that it might be wai's flagship product in the future, so he discussed the name with Guster. For the new brain of www.piaotia.com. This is exciting research, and the project team includes nearly two thousand people from the Artificial Intelligence Research Institute.
If you want to reach the 8.0 threshold in theoretical derivation, you must break through the "plateau barrier". This is a bit troublesome and requires new ideas to improve algorithms and data structures. The standard artificial neural network algorithm used by wai was jointly developed by Wang Chengmin and Guster. However, the research of the brothers has different emphasis, and many people have surpassed Li Zhuang's level in their own fields. For example, Jing Ran's algorithm Achievement. Weibel studied knowledge processing and symbol processing systems. Although he did not reach the level of his teacher, he was more comprehensive. He is only 22 years old this year, and it may not be impossible to surpass his teacher in the future. Compared with him, Zhen Yun is very good at parallel computing network and image search, especially image search. Maybe it will bring him a trophy.
??The former computer-related employees of the Institute of Applied Physics have been transferred to the Artificial Intelligence Research Institute. The project team led by Wang Wencai and Emerson is developing the low-power hardware required for the 100m intelligent brain, and the workload is not small. Liang Can Laoji kept the factory affiliated with the institute in good order and continuously provided the components needed for the 100m project.
Bernard could not escape his senior brother's assignment. His task was to design a power management system for 100m. The total power should be controlled at 15,000 kilowatts. The voltage and current compensation and suppression system must be very complete, and there must be no serious problems, otherwise There will be heavy losses. The hardware cost of 100m is as high as 20 million gold coupons, the operating conditions are complex, and the importance of power management is mentioned to the highest level.
R&D follows Lizhuang¡¯s habit of solving problems step by step, which not only makes it easy to verify ideas, but also commercializes the results generated during the R&D process, and the profits are used to support the next step of research.
The cost of hardware is far higher than expected, mainly due to the terrible energy consumption and the cost of long-life components. Calculated based on medium-power operation of 10,000 kilowatts, 100m consumes 87.6 million kilowatt-hours of electricity per year. In addition to various losses and subsequent maintenance, attention must be paid to it. Integrated chips cannot be used at the beginning because they are verification machines and are immature. The non-integrated mode is convenient for maintenance and improvement.
The tentative plan for 100m is to highly imitate the human brain, and at the same time make changes based on the characteristics of the silicon optical network, using three models plus a smart ring. Intelligent brain is essentially the research that combines the creative thinking of the human brain with the high-speed parallel processing capabilities of traditional computers, taking into account the advantages of both and trying to avoid the disadvantages.
Power consumption is a flaw in smart brains. Al-level aerospace speakers use a large number of integrated chips.The consumption is still 3~10kw, and the activity of the human brain is only equivalent to 20w. The human brain accounts for an average of 2% of the total body weight, but uses 25% of oxygen, compared with only 12% for the kidneys and 7% for the heart. The transmission speed of nerve signals in nerve or muscle fibers can reach 370 kilometers per hour, which is slightly exceeded by people with special training. The company's biological robot is called a biochemical robot after strengthening. Its bone density, muscle density and flexibility have reached a very high level. It only takes 4.5 seconds to run 100 meters.
There are many mysteries in the human body. The total length of the neural network exceeds 70 kilometers, and the nerve cell circuits are nearly 1,500 times more complex than the world's telephone network. The human brain is carrying out 100,000 different chemical reactions every second. The fastest nerve impulse transmission speed between nerve cells exceeds 400 kilometers per hour, and can process 86 million pieces of information every day.
What is astonishing is the capacity of the human brain. Less than 10% of the more than 10 billion nerve cells are developed and used, and it can store 500 million copies. The power of the astronomical network is fully reflected in the biological servants. The Human Gene Research Institute found The intelligence level of biological servants using hybrid chips is not inferior to that of V-level intelligent brains. If it were not for the limitations of the chip, they could be smarter, but of course the stability would be reduced.
The 100m level is part of the infinitely expandable carbon-silicon optical network brain, which is easy to maintain and replace. Compared with it, the biological computer with unique advantages is another major project of the Artificial Intelligence Research Institute.
In the later stage of the adjustment, the Biochip Laboratory of the Institute of Human Genetics was placed under the Institute of Artificial Intelligence and renamed the Institute of Biocomputing, and elite forces were recruited from the Department of Biology and related institutions to join. The quirky Brain Research Institute was officially disbanded, and its subordinate institutions were placed under the Theoretical Research Institute, including the Soul Laboratory, which was considered neurotic by senior officials.
Emma DuBois was re-elected as the director of the Institute of Biocomputing, Wan Weichen, the former director of the Biochip Laboratory, was appointed as the deputy director, and Glenn Bick, a Dutchman from the Institute of Nanotechnology, also became the deputy director aggrievedly. This girl is so solid that Wan Weichen, who has been tested for a long time, gave up after the competition. Besides, he is not good at overall design. Dr. Bick is an expert in DNA computers and cell biocomputers. His management skills are not very strong, so he has no intention of competing for the top spot. Teams of young people leading middle-aged people are very common in research institutes. Everyone is accustomed to it. Many old researchers are praying not to recruit too many top experts. The gap in ability will make them lose face.
Although the company¡¯s research in the field of biocomputers is late, it has made many achievements. However, it has never systematically researched biocomputers. This time it is considered to be entering the civilian field from the military industry, and the starting point is very high.
Biochip is an integrated circuit made by scientists using the switching properties of proteins and using protein molecules as components. Computers using biological chips are called protein computers or biological computers.
The biochips currently developed by the company are biomemory chips and biocpu. The auxiliary chips have not been developed before. They all use the nerves and cell tissues of the human brain to control the body. There are far more types of biochips, such as synthetic protein chips, genetic generation chips, and hemoglobin chips, etc., which are relatively common. The company has purchased similar products on the market.
Theoretically, a chip made of protein can accommodate hundreds of millions of circuits in an area of ??1 square millimeter. Because each of its storage points is only the size of a molecule, its storage capacity far exceeds that of an ordinary computer. On the evaluation form of the Artificial Intelligence Research Institute, biological memory chips are technologies that are tied to atomic hard drives. Each has its own advantages and disadvantages. The capacity is larger, but it is not as stable as the atomic hard drive, and has limited use conditions and a short lifespan.
Ultra-density optical storage is the ultimate goal of research and development. There are many difficulties in this field, and perhaps there are problems with the theoretical derivation itself. Xiong Fengsheng, director of the Optical Storage Laboratory, was surprised by Li Zhuang's confidence. He didn't ask too many questions. His boss supported him and he was in his own field. He felt comfortable doing research.
Biochips have been a hot topic in the world in the past two years. They have low impedance and low energy consumption when transmitting information. They also have biological characteristics and have the functions of self-organization and self-repair. It can be combined with the human body and the human brain, obey the instructions of the human brain, and absorb nutrients from the human body. Integrated circuits made of proteins are only one hundred thousandth of silicon integrated circuits. They can operate at a speed higher than that of the human brain. The component density is 1 million times higher than the density of brain neurons. The speed of transmitting information can be imagined.
When scientists used to describe the prospects of this kind of computer, they made many points, such as using it in medicine to restore sight to the blind, greatly improving the memory of the human brain, enhancing physical fitness, etc.
The first biological computer with complete significance was made by Citizen scientists five years ago. Newborn babies were fragile and made a joke. However, R&D institutions continued to work hard and the level improved rapidly. However, the sales volume was very small and maintenance was troublesome. Biological servants using hybrid chips work stably. Although they are not mass consumer products, ordinary people still see the potential of biological combat robots.
?Biological computers have just appeared.It was considered to be a crisis by many media and was specially approved in large numbers because it could lead to the control of ordinary people and make life worse than death. Many people believe that if an artificial biological computer is implanted into the brain, once a malfunction occurs, it will greatly affect the operation of the brain, ranging from paralysis to death. Its power consumption and stability are not something ordinary people can figure out. Compared with the many claimed advantages, the military potential is greater, and the stability of human society has more variables. In the eyes of many people, when biotechnology reaches its peak, it may be the end of human society. They have called in the media for a ban on biocomputer research because their many functions can be solved by other means.
The use of carbon nanotubes to replace silicon as the semiconductor material in chips has been researched for a long time, and now it can achieve ultra-low energy consumption. Among the three types of computers in cutting-edge research, namely molecular, photon and quantum computers, biomolecular computers are the most mature and most embarrassing. The advantages and disadvantages are equally obvious, because humans are carbon-based organisms.
Molecular, photon and quantum computers can all be found in related research institutions at United Technologies Corporation. The Photon Computer Laboratory is a subsidiary unit of the Intelligent Machine Research Institute, and the Quantum Computer Laboratory has just been transferred from the Department of Physics to the Artificial Intelligence Research Institute, plus the biological computer The Research Institute and the Artificial Intelligence Research Institute will bring all relevant research under its umbrella and move towards higher goals.
With the progress of cutting-edge research, researchers are becoming more and more worried, and the Ministry of Security Research Center is considering countermeasures. Unpredictable voyeurism is becoming a reality, such as that weird quantum ratchet. It uses an oscillating signal or a randomly changing signal to obtain useful movement that can control the direction from a seemingly chaotic state, allowing electrons to jump from one electrical component to another without wire connection, which is a bit like a space jump. Network physical isolation now needs to be re-researched, and it cannot be solved by disconnecting the hardware.
In the Artificial Intelligence Research Institute, the Smart Machine Research Institute is the core department, and the others are auxiliary institutions, whether it is molecular and photonic computers with low power consumption and high efficiency, or quantum computers that temporarily consume power and can scare people into turning pale. They are all auxiliary tools of Intelligent Brain, but Intelligent Brain itself uses a variety of technologies, and they are a mixture of different emphases.
The 100m in the design has a complex hardware architecture and uses three types of components: carbon, silicon and light. There are nearly forty versions of component combination schemes, and some of the design concepts of quantum computers are also incorporated.
The workload of the software architecture is greater. The algorithms derived from preliminary research include standard artificial neural network algorithms, genetic algorithms, chaotic search and the Nebula algorithm developed by Jingran. Different hardware and software are connected using intelligent rings, and the data structure is a tree. Shape and mesh structure to minimize the trouble of information retrieval and the resulting data redundancy.
The 1000m-level design is in planning, which is a linearly amplified version of the 100m-level. If the 100m-level can be developed, there should be no problem for the 1000m-level to reach the level of adult intelligence.
¡°Boss, Sunway and the others have finished reading.¡± Arthur poked the petrified computer expert and reminded him.
"Oh, then go home. There's actually nothing to see here," Li Zhuang continued, "It's boring to follow me. If you don't want to do this job, you can apply to be transferred to another department."
Arthur shook his head, hesitated for a moment and said, "I can only be a bodyguard. It's the same everywhere. Besides, there aren't many reliable people around you. The important thing is that being your bodyguard is a small workload, haha!"
Li Zhuang smiled and was about to speak when Shuangwei rushed over and said, "Dad, I want to raise a koala!"
; - Didn't finish reading? Add this book to your favoritesI'm a member and bookmarked this chapterCopy the address of this book and recommend it to your friends for pointsChapter error? Click here to report
- Previous pageCatalogNext page