column opening language
This is an era full of tests and changes,
This is an era of a closely connected community of destiny.
In this cold winter when the epidemic continues to rage and keep warm;
On the eve of frequent extreme weather and drastic climate change;
In this virtual reality tide, surging moment;
We need the power of science and technology and the power of reason,
We are convinced that this is the only way,
We can break through the prison of reality,
To explore the wider world,
To pursue a fuller self.
The bows and arrows of this era have opened,
The most vigorous arrow of science and technology will fly out!
In 2022, we launched the new year's great contribution,
Touch the cutting edge of technology,
Let's see the future together!
For a long time, supercomputers have always been a popular element in science fiction -- supercomputing in different forms and highly artificial intelligence in terminator, gospel warrior of the new century and super body, which carries human imagination about the quantum world and the potential of human brain.
Not long ago, the epic science fiction novel dune created by Frank Herbert was adapted into a film of the same name, which was popular all over the world. The film presents a desert planet Arrakis with rich details, which makes people feel like they are on the scene.
A group of British scientists who study climate models are deeply attracted by the strange planet and try to simulate the climate environment of erakos. However, the complicated calculation of Beijing Vastdata Technology Co.Ltd(603138) such as clouds, temperature, precipitation and wind force can not be done manually. So the scientists called a supercomputer to process data day and night, build models, and finally come to the climate environment envisaged in Herbert's book - which is largely consistent with the real climate model.
visual demonstration diagram of erakos climate model picture source: climate archive
Obviously, the above is only the "tip of the iceberg" of supercomputing power. From paleoclimate research, earthquake early warning, geological exploration, and then to stellar trajectory deduction, supercomputing has shown magical power in many fields. While continuously promoting industrial reform, supercomputing continues to expand the boundary of human exploration. So, after decades of development iterations, what kind of development form will supercomputers show in the future? What are the frontier research fields of brain computer fusion? What are the weaknesses in the development of China supercomputing?
To this end, the reporter of the daily economic news had an exclusive dialogue with Steve Chen, an academician of the National Academy of engineering, an academician of the American Academy of Arts and Sciences and a world-renowned supercomputer pioneer. At the same time, Chen Shiqing is also the founder and President of the third Brain Research Institute of the United States and the chief scientist of the National Engineering Laboratory of big data application technology of China Aerospace cloud network industry.
In the 1980s and 1990s, Chen Shiqing was regarded as a "superstar" in the field of supercomputing in the world, and Time magazine once took him as the cover character. He learned from David kuck, the chief system designer of NASA's early supercomputer iliac-iv, and joined the famous Cray research in 1979 as the chief designer and senior vice president. Chen Shiqing photo source: photographed by reporter Zhang Jian
As a world-famous supercomputer expert, Chen Shiqing achieved four "world firsts": in 1983, he developed the world's first 2-cpu parallel vector supercomputer (Cray XMP / 2), which laid the foundation of the system architecture of parallel vector supercomputer in the world; In 1998, it developed the world's first brand-new supercomputer system architecture and application concept - blade supercomputers; The world's first information network based on supercomputer and global networking to form a current grid; The world's first Internet-based, application to application, dynamic, low-cost and efficient enterprise collaborative operation intermediate software tonbu has greatly promoted the development of global supercomputing.
"The server of blade supercomputer is like a 'blade', which can be inserted into a rack chassis. A 'blade' is a system motherboard. It can not only run its own system independently to serve different user groups, but also integrate into a service group, share resources and run a single task assigned. It has high parallelism, real-time cooperation, low power consumption, low price and small space And so on. " Chen Shiqing told the reporter of the daily economic news.
When talking about the development trend of supercomputing in the future, Chen Shiqing told reporters, "today's supercomputers have reached a development bottleneck due to too much energy consumption. In the future, supercomputing will evolve from centralized to distributed, and go deep into communities, streets, industrial parks, universities and research institutions. At the same time, supercomputing and brain science will also move towards deeper integration."
"supercomputing in the future should be distributed and the world should become a network"
NBD: at present, the mainstream supercomputers basically adopt the centralized computing power mode. What are the shortcomings behind the realization of super computing power output? What will be the evolution of supercomputing in the future?
Chen Shiqing: from the 1980s to the 1990s, I was only doing one thing - inventing and manufacturing the fastest supercomputer in the world. Based on the system architecture I designed, the centralized supercomputer began to develop rapidly and is widely used in national defense, petroleum, aerospace, aviation, automobile, nuclear power, meteorology, chemical industry, materials, biology, pharmacy and other fields. For example, at & T, Boeing, general motors and Mercedes Benz are all my customers.
Among all the large-scale supercomputers in the world, 80-90% used the system architecture and concept I created at that time, from single machine to dual machine, one to two, two to four, four to eight, and evolved into thousands of parallel computing.
But with the increasing size and speed of supercomputers, their energy consumption is also higher and higher. At present, the top ten supercomputers in the world need a power supply station with a scale of at least 20 megawatts to support them. The power consumption per hour alone exceeds 20000 kwh, which is equivalent to the daily consumption of a small town.
Although I created the centralized supercomputer architecture, the problem of huge energy consumption has never been overcome. Without the financial support of the state or large enterprises, this model is difficult to be sustained and widely applied to the new scenario of digital economy in the future. From the national level, this large centralized supercomputer still has a huge demand in the fields of lunar exploration, meteorological monitoring, oil exploration and so on. Due to its huge energy consumption, many small and medium-sized enterprises are difficult to bear in the market.
So at the beginning of this century, I began to rethink another system architecture - from centralized to distributed. Its core purpose is to make supercomputers faster and consume less energy. This is a completely different field.
NBD: how to understand that all supercomputers in the future should be a huge "distributed network"?
Chen Shiqing: I think all supercomputing in the future should be distributed, supplemented by artificial intelligence, big data and other technologies, and the world will become a network. The edge / bottom layer of this network is communities, streets, industrial parks, etc., where there is big data for all applications, and supercomputing can be accessed at any time. This is a new idea.
At present, we are still in the pilot stage of developing network format intelligent supercomputing. For example, if a city is divided into 5000 or 50000 grids, a small supercomputer can be set up in each street or community, and all local data can be calculated, transformed and served in real time. Then all communities are combined to form a huge distributed grid. Similar to the current power grid, in the supercomputing distributed network, the future computing power, capacity and algorithm are also distributed and can be complementary.
For example, we can upload the big data of each community in the bottom network format intelligent supercomputing designed by Jiading, Shanghai; You can also build a network format intelligent supercomputing in the West Lake District of Hangzhou, digitize all urban changes, deeply mine and integrate the data on the same platform, and then use it for intelligent medical treatment, education, transportation, manufacturing, cultural innovation, energy, finance, environmental protection, security and other terminal services.
NBD: as a top scientist spanning many times, what changes have you observed in the concept of human-computer interaction in the development process from minicomputers, supercomputers to supercomputers to cloud computers?
Chen Shiqing: the concept of human-computer interaction has evolved for a long time. The earliest human-computer interaction was centralized and mechanical. Throw everything there and let supercomputing help. Nowadays, the concept of human-computer interaction has changed from mechanical and centralized to a more intelligent and personalized mode.
The future supercomputing should be closer to you and more personalized. For example, where you live now, there is a small distributed intelligent supercomputer next to your community. The data it needs comes from this community, and its complete application scenario is also in this community, and serves the people of the whole community.
If one day, a heart patient suddenly falls on the road. How do we immediately know the type of heart disease in this patient? Which hospital has a specialist in this kind of heart disease and is nearest? How to dispatch the rescue force as soon as possible? These require the use of intelligent supercomputing to quickly calculate all data, including patient status, positioning, road conditions, hospital conditions, etc. This is one of the important directions and application scenarios for the development of intelligent supercomputing in the future.
"supercomputing should develop towards brain like direction in the future"
NBD: in addition to the trend of distributed and grid, how will supercomputers evolve iteratively in the future?
Chen Shiqing: I think the future supercomputer should also develop in the direction of "brain like", with small volume, fast operation speed and low energy consumption.
In fact, the daily consumption of the human brain is only 25 watts, but it can command about 100 billion neurons and process a huge amount of information every day. Human brain, as the most delicate intelligent organ of human body, its "performance" is stronger than supercomputer. At present, the development of human brain is only 5%. If we can improve the development of human brain with the help of brain science and combine it with supercomputing, we will create valuable application scenarios.
NBD: you prepared to establish the third brain research institute in Silicon Valley in 2013 and officially established it in 2019 to focus on the research of brain computer integration. In your opinion, what is the core idea of the "third brain"? What are the main research directions around the "third brain"?
Chen Shiqing: in fact, more than a decade ago, my research direction turned to brain like computing. I hope to rely on "elastic distribution" to make supercomputers as low energy consumption and high efficiency as human brain, and improve the development degree of brain at the same time. For me, such a turn is tantamount to a "self revolution".
In addition to the brain (first brain) and cerebellum (second brain), each of us has an "extended" third brain. The "third brain", also known as bionic brain, is a comprehensive science combining cognitive neuroscience, biological information science, biological system science, bioengineering science, life science, clinical medicine, information engineering science and supercomputer science. Relying on the highly trusted, secure, efficient, distributed and collaborative super grid cloud platform, we have carried out R & D work all over the world.
At present, the "third brain" mainly has four research directions - brain research, brain protection, brain development and brain extension. The research in the field of developing brain is relatively mature.
The first direction is to study the brain and deeply understand the basic biological mechanism of the brain. Only after we understand its basic architecture and why it is intelligent can we carry out brain like computing. Imagine that the brain commands about 100 billion neurons and consumes only 25 watts, half of which is just for the biological needs of the brain. Such a structure is full of mysteries.
The second direction is to protect the brain. After in-depth research on the brain, we will know what its shortcomings and weaknesses are, so as to know how to take protective measures to prevent brain injury and find early and accurate detection and intervention methods for brain diseases. Our research focuses on how to carry out early detection, predict and intervene in advance from the information level, physical level and chemical level, reduce the incidence risk and move forward the "window period" of treatment.
The third direction is to develop brain, research and develop efficient learning methods of brain computer integration. God has given us such a smart brain. We should focus on how to give full play to one's brain potential without being suppressed by other factors.
We conducted a pilot in Fremont high school in the United States, and finally proved that the "third brain" related technology can significantly improve the learning speed and effect. I hope that in the future, we can use our theories and technologies to fully stimulate the learning potential of the brain. We can finish kindergarten at the age of 17. After three years of social practice, a person can have the thinking and analysis ability at the doctor level at the age of 20. "Although I am 77 years old, my brain is only in my 30s. My thinking and innovation ability is comparable to that in my 30s."
The last direction is to extend the brain, extend the brain computer integration technology to the cloud, and realize lifelong companionship from childhood to old.
NBD: now, research in the field of brain science has long been on the cusp of the science and technology circle. Tesla CEO Elon Musk and Google founder Sergey Brin have joined one after another. So, what are the cutting-edge application scenarios of brain science?
Chen Shiqing: taking the field of intelligent medicine as an example, we are working with world-famous medical colleges to study brain diseases and cancer prevention, realize "unpredictability" with the help of brain science and artificial intelligence, diagnose diseases in advance and realize non-invasive disease treatment at an early stage.
For example, according to the research of brain visual nerve, before Alzheimer's patients begin to lose memory, use the newly developed precision instrument to detect the decline of brain cognitive function early, and carry out non drug physical intervention and rehabilitation to improve personal quality of life. At present, Parkinson's disease, autism, depression, Alzheimer's disease and epilepsy are the key projects of intelligent medicine.
In addition, taking intelligent education as an example, we hope to find a law from brain science, which can judge whether a child is suitable for scientific, technical, engineering, artistic or medical learning and education very early.
Brain science combined with artificial intelligence will transform the future education system. China's first demographic dividend has passed. In the past, we relied on labor-intensive. Now the society has entered the aging stage and needs to rely on the power of science and technology. Brain science can accelerate the release of young talents to society and help to realize the second demographic dividend.
NBD: previously, when talking about brain computer integration, Elon Musk thought that human beings had only one choice - to become artificial intelligence (AI). What do you think of this statement?
Chen Shiqing: in fact, musk sometimes expresses a little extreme. The brain computer integration and brain computer interface technology he talked about are the combination of brain science and AI technology to help disabled people live a normal life by implanting chips.
But I think another important development direction of brain science is to stimulate the brain potential of most people and promote thinking and innovation. The human brain is the smartest. We can always think and invent a new tool to enhance our ability in some aspects, just like telescopes and microscopes to the human eye, cars and planes to the legs. Artificial intelligence is only a part of human "extension". Artificial intelligence alone is not enough. It must be deeply combined with brain science.
In fact, the best human-computer interaction in the future is to hand over the heavy and complicated work to artificial intelligence, and then let our brain focus on thinking and innovation through brain computer integration technology.
"China's supercomputing has reached the world leading level, but there is a shortage of Applied Talents" Chen Shiqing picture source: every reporter Zhang Jianshe
NBD: over the years, your scientific research experience has been inextricably linked with China, which has promoted the development of China's medical and health big data platform. Please share your experience in participating in supercomputing projects in China.
Chen Shiqing: as early as 2004, I decided to return to China and initially designed and completed my third generation super blade computer in Shenzhen. At that time, it was not sold for the first time, but brought a new generation of supercomputing to China's poor and remote rural and ethnic minority areas.
We have set up four clinical pilots in these areas, including Kangbao County, Zhangjiakou, Hebei, Xilin Gol League, Inner Mongolia, Kazak Autonomous County, Aksai, Gansu, etc. My team and I spent seven years in-depth local research, collected and mined a large amount of medical data, especially the local big data of chronic diseases, successfully established an information grid model of chronic diseases, realized the interconnection of basic medical services at the county, township and village levels, and formed the version 1.0 of the "medical Consortium" of China's chronic disease medical system today, It also promoted the development and improvement of the early rural "new rural cooperative medical payment system.
In 2007, Gao Qiang, then Minister of health, visited Kangbao County, Zhangjiakou. He was very satisfied with our medical data and grid model. "Great, this is China's future big health platform." Later, under the promotion of Minister Gao Qiang, 200 counties received active funding from the national medical reform fund, and the medical big data grid model was promoted in more places. Later, it gradually became a national chronic disease management platform.
From 2006 to 2012, my whole focus of work shifted to Beijing, formed a team with independent financing, continued to develop the 4th, 5th, 6th and 7th generation blade supercomputers, developed hospital medical software, and established the earliest health cloud computing and big data service platform in China.
At first, our idea of version 1.0 was to combine counties, townships and villages to create a "Medical Union". In the future, we hope that all China will be a network and use a health code. No matter where your family history is, it can always be transferred in time.
NBD: in recent years, China has gradually occupied the commanding height in the world in the number of TOP500 high-performance supercomputers, and supercomputing represented by "Shenwei · light of Taihu Lake" and "Tianhe 2" has repeatedly led the world in computing power. According to your observation, what level is China's supercomputer in the world?
Chen Shiqing: in fact, from the perspective of system design and manufacturing, China is already at the cutting-edge level. Many of the top ten supercomputers in the world come from China. For example, at the global supercomputing conference not long ago, the Chinese supercomputing application team led by Zhijiang laboratory won the "Gordon bell" award in 2021, the highest international award in the field of supercomputing application, by virtue of the quantum computing simulator developed by the new generation of Shenwei supercomputer - "real time simulation of very large scale quantum random circuits" (swqsim). This is one of the epitomes of the development level of supercomputing in China.
NBD: so, what are the shortcomings in the development of supercomputing in China?
Chen Shiqing: first, there is a shortage of supercomputing applied talents in China. The talents trained in China are concentrated in large institutions such as the Chinese Academy of Sciences, which mainly study centralized supercomputing. The field is relatively narrow, and only national institutions can afford it.
In the future, distributed intelligent supercomputing will be small supercomputing scattered in communities, streets and industrial parks, and there will be more and more subdivided and complex application scenarios. In view of this, there is still a big gap in the application-oriented talents of artificial intelligence in China. At present, at least 5 million relevant development talents are needed. Previously, we established an artificial intelligence college in Chengdu Jincheng University, hoping to cultivate more medium and high-end application development talents, and plan to copy the successful training model to 100 universities.
In addition, when the market talks about supercomputing, it talks more about specific industries and products, and few people talk about computing power and capacity. In fact, the development of intelligent supercomputing in China is a big gap in computing power and capacity. In the future, all artificial intelligence scenarios need algorithms, and algorithms need computational support, which needs to be vigorously developed.
At present, some cities realize that the core of the development of digital economy represented by artificial intelligence lies in computing power, and have begun relevant layout. Huawei, Inspur and other enterprises are also participating in the construction of urban computing power and capacity. China should focus on industrial parks and communities, lay out new infrastructure and computing power such as 5g base stations, and create a good supporting environment for the innovative development of small and medium-sized enterprises.
Finally, only by combining brain science, artificial intelligence and intelligent supercomputing can we complete the application scenario from end to edge and establish a complete industrial chain.
NBD: last question, what have you experienced from clay company, which initially monopolized the world supercomputer market, to parting ways with its founder Seymour clay? As a top scientist and a senior entrepreneur, how do you keep a balance between technology exploration and business?
Chen Shiqing: because the supercomputing system he studied was surpassed by me (laugh). At that time, my second generation supercomputing was developed, but his second generation has not yet come out. Slowly, the whole company's revenue "depends on my products".
In fact, every generation of supercomputing costs a lot of money to support the R & D team of hundreds of people. At that time, the company could not support two R & D teams that were "burning money". On the one hand, clay is the founder after all, and the company must give priority to supporting his ideas and research; On the other hand, I have brought a lot of commercial benefits to the company, but I still can't raise enough follow-up R & D funds. Over time, we "separated", and then I took the initiative to leave the company with a research team of 45 engineers.
All along, I have been exploring more frontier fields. My supercomputing can certainly create considerable benefits, but I put it into the next research. Research - new product launch - commercial realization - re research, such a cycle.
My outlook on life is to benefit more Lbx Pharmacy Chain Joint Stock Company(603883) from my scientific research achievements. Whether Chinese or Americans, the world needs help. It's good to do this! I hope people engaged in science and technology industry can realize their ideas and values. The purpose of doing scientific research is not to make money, but to make technological dividends benefit all people and help people at the bottom of society solve difficulties. The way of science and technology, with the rain!
(Daily Economic News)
