Then, on the last day of 2019, the first thing mentioned in aqin 3.0 was started: he was employed as a chair professor of intelligent science by Tsinghua University, and led the establishment of Tsinghua Institute of intelligent industry (air). Facing the fourth industrial revolution, it aimed to create the worlds top innovation and research platform by taking automatic driving, artificial intelligence + Internet of things and brain like intelligence as the key technology breakthroughs.
Then, his old friends, Ma Weiying and Zhao Feng, joined the team one after another.
But all this is not without curiosity.
Retirement at the age of 53 is just the age of youth. It also coincides with the explosion cycle of new technologies: AI, 5g, autonomous driving, biological computing, quantum computing Is he really willing to retreat? What are the opportunities behind joining Tsinghua? Why is the Intelligent Industry Research Institute established? How will it become bigger and stronger? How to persuade Daniel such as Ma Weiying and Zhao Feng?
And whether the epidemic has disrupted the established rhythm, and what advice does he have for young people under the challenge of the times?
So this time, we asked Tsinghua to teach, talked about technology trends, listened to the thinking under the epidemic situation, and even got the secret of family education shared by Yaqin.
This is also Zhang Yaqin announced his retirement, the first interview this year to answer all the concerns.
Yaqin said that after 20 years of study and 30 years of work, he has been thinking about how to repay the country and society by cultivating talents and driving industries. This is his deep feelings.
Seeing that the development of science and technology is ushering in a golden age and an unprecedented period of technological explosion, he thinks it is time for all.
However, all these topics still need to start from the day of 3.0 official publicity.
Retirement from Baidu, in fact, has been planned for a long time.
Ma Weiying and Zhao Feng join? After sharing the plan, it is easy to get along without Persuasion.
Any research institutions, once used as indicators, will start to deviate.
We are now in a technological cycle in which digital intelligence, physical intelligence and biological intelligence blend.
Three things that happened a hundred years ago are being repeated in other ways.
Im not a tiger dad. I dont criticize children. Family education lies in trust and company.
Qubit: you retired from Baidu, which makes people feel very sudden.
Zhang Yaqin: maybe it just seems sudden, but actually I have planned for a long time.
Robin invited me to join Baidu in 2014. We have been friends for many years, and out of recognition of Baidus technology belief, I agreed.
Finally, I retired from Baidu for five years and three weeks.
Qubit: when do you really think?
Zhang Yaqin: about the end of 2018, at that time, the new technology and business had reached a certain level, and I could leave.
Including autopilot, cloud computing, chip, quantum computing and other new technologies and services, layout and establishment are completed, all of which are a time when I think we can take a step away.
And, crucially, the new leaders are up to take over.
Let me want to realize the feelings I always have.
Qubit: cant be expanded without retirement?
Zhang Yaqin: in the life of 3.0, I dont want to spend too much time on operation.
In Baidu or Microsoft, the companys business operations are very time-consuming.
In the first line, we should pay attention to the operation of Baidu, especially its products.
Qubit: so the internal handover started very early?
Zhang Yaqin: at the end of 2018, I started to discuss with Robin.
Therefore, the internal communication, organizational restructuring and external announcement were carried out in an orderly manner according to the plan.
The overall handover took about half a year, and the process was not so sudden.
Qubit: including later retired, baidu Apollos cooperation is also promoting.
Zhang Yaqin: Yes. I see Baidu products, still very emotional, with the Microsoft products. And Baidu colleagues also often meet. Im still chairman of the Apollo alliance, and Im still working closely with Baidus team. In the past two days, Im still trying to take the Apollo go unmanned car in Beijing.
So retirement itself, its not that Im not satisfied with Baidu, or Im not satisfied with my job.
We have such a long life. We have to have different priorities at each stage. I hope that we can do one thing again.
Qubit: Baidu is facing some disputes in recent years, and your retirement happens to be at this time point, which makes many people feel sorry.
Zhang Yaqin: there may be some different opinions on Baidu, but Im still optimistic.
I think the market has underestimated Baidu. Baidu is still a leading technology company.
Zhang Yaqin: I hope to build an Intelligent Industry Research Institute for a long time. I have to work closely with universities. There are three main reasons for choosing Tsinghua University.
First, if the institution is in Beijing, it will be relatively convenient for both university scientific research resources and my personal life radius.
The second important factor is the sincerity from the president to the Department of Tsinghua University.
Qubit: thirst for talents?
President Qiu really expressed his infinite sincerity. He particularly supported my idea of building an Intelligent Industry Research Institute. He said that he hoped that you could come and do whatever you want to do, and better achieve the goal based on Tsinghuas platform, resources and global influence. And then they talked about it at least five times. Yang diange, President of vehicle and Transportation College, also contacted me many times. In fact, the first time he contacted me was diange. He saw my retirement report that day and wrote me a long wechat, which moved me very much.
Qubit: also have the best scientific research resources?
Zhang Yaqin: Yes, this is the third point
Artificial intelligence needs interdisciplinary thinking and cooperation.
Qubit: all kinds of resources are very good.
Zhang Yaqin: the advantage of Tsinghua University is that these disciplines are in the leading position, with high level, and there are many famous scholars.
The students are also excellent. In the last two weeks, I just recruited three doctoral students, one of them is better than the other. And this year, the best PhD students are in school.
Qubit: you just said that you always have a feeling in your heart?
Qubit: now this matter is specifically Tsinghua Institute of intelligent industry?
As mentioned before, the main purpose is to provide technological driving force for the fourth industrial revolution. Therefore, to be a global and intelligent industrial research institution is very focused and its positioning is very clear from the beginning.
Qubit: why is this positioning?
Zhang Yaqin: actually, Ive been thinking about this for a long time.
I used to work at Microsoft Research Asia (MSRA) and earlier at Sanford Research Institute (now SRI). The models have their own characteristics.
The positioning of MsrA is very clear. It is application research. Its main customers are the product departments of Microsoft. Therefore, there is such a problem that the final technology is limited by the productization and the companys strategy.
SRI is a social research institute that has produced many cutting-edge technologies and companies, including Siris parent company, which Apple acquired later.
But the problem with Sri is that it does not have deep binding with a school, so its stability and continuity are not enough.
So now Im doing air (Tsinghua Institute of intelligent industry) to sum up and synthesize the experience of Sri and MsrA mode. It needs not only the depth of University binding, but also the flexibility of facing the whole industry.
To tell you the truth, such talents cant be told through lectures. It needs leading talents with such experience to cultivate.
Ma Weiying and Zhao Feng, who have been promoted by the government, are such leading talents. They have both profound academic attainments and senior industrial background.
What you think an organization can be made of depends on what the first few people look like.
Even if he doesnt need to tell you which direction he is going to go, his usual thinking and scientific research methods will naturally promote him to go in the right direction.
Qubit: is there roadmap?
Zhang Yaqin: Yes, but I dont want to share it now.
I dont want to define this framework at the beginning. I hope to do it very well, not too many people.
There are three directions I have set, and the chief scientist will lead the team in each direction.
There are some researchers, some doctoral students, maybe postdoctoral students and people engaged in industrial transformation in this team, so each team is not very large. I think it should not exceed 150 people if we add one piece, and we should never exceed 200 people.
Qubit: and how to make it bigger?
Zhang Yaqin: it must depend on the transformation of achievements. For example, if there are achievements, they can be licensed to other companies, and authorized technologies can be transformed into products and enter the market through cooperative companies.
Another more direct way of transformation is that we have a very core technology. Investment into a start-up company must be very focused in air, focusing on scientific research.
Tsinghua University, scientific research personnel and air can take shares in the companies whose core technological achievements are transferred out, and the CTO and chief architects trained can also come out.
Therefore, air itself will not be very big. If air wants to be refined and strong, the products and companies transformed from technical achievements can be very strong and produce great value.
Qubit: is there a ready reference?
Zhang Yaqin: now I cant find the reference completely. MIT multimedia lab? But they are still more basic research.
You know what? The number of papers published by Microsoft Asia Research Institute is very high, but we have never asked for it, and we will not say that you have published 5 papers more than 1 article.
What needs to be seen is the contribution to technological products and society.
Also to air, I would rather you do a great contribution to the society, rather than do a hundred small things, sewing and mending, see the number of papers and patents.
Any organization, once used as an indicator, will start to deviate.
I want to do this thing, the number is not the goal, not the goal, but the natural result of doing a good job.
Qubit: are all doctoral candidates recruited?
Zhang Yaqin: they are all doctoral students.
Zhang Yaqin: it is to be excellent. However, in terms of specific aspects, I hope TA is interested in practical problems. Although it is OK for doctoral students to do theoretical research, I prefer industrialization in several directions.
Qubit: do you need strong persuasion for big cattle like Ma Weiying and Zhao Feng to join?
Zhang Yaqin: when I wanted to do this, I made a list of them, including those who would like to join in the future. In fact, Im just gathering people with common beliefs.
Their ideas are very similar to mine, so Persuasion is more about sharing and thinking about this matter. Because we all share the same aspiration, we can get along with each other.
Qubit: are these big bulls also the paradigm of talents to be cultivated in the future?
In my opinion, this is also the most needed talents in China. I hope that the talents trained in the future will have such qualities.
They can have a systematic thinking and can design the overall and underlying technical architecture and product architecture.
The gap between us and the United States is also reflected in this respect. How many years has Microsoft used a set of architecture, whether operating system, cloud computing or database? Google, too, has a big platform, big architecture, and large-scale thinking.
Qubit: three specific directions of Air Research Institute, also the result of careful consideration?
Zhang Yaqin: there are three aspects to consider
First of all, I hope to select some areas with huge business space and development opportunities.
Thats how cars and transportation are. Its a big industry. AI + Internet of things, which involves the industrial Internet of things, intelligent manufacturing and robots, is also a large industrial field. AI + biology, including big health and healthcare. And these have a great impact on Chinas entire industrial transformation.
Second, AI plays a particularly important role, even a decisive and subversive role. In the four modernizations of automobiles, intelligence is the decisive factor. AI also has a huge driving force in the research and development of new drugs involved in medical treatment.
Third, huge social benefits. Automatic driving can reduce traffic accidents, and medical treatment is closely related to everyone. Therefore, we hope that the industrial value and social value can be unified.
Finally, there is one personal interest. I am personally interested in autonomous driving. Ma Weiying has a passion for medical care and education
Qubit: also in line with technological evolution and historical development trend?
Zhang Yaqin: on the general trend, I think these three industries will continue to affect the next 50 years.
I think the technology cycle we are in now is the integration of three intelligences.
Digital intelligence, physical intelligence and biological intelligence.
Digital intelligence is the promotion of complete digitization; the digitization of the physical world, which I have talked about for many years, is directed at this by industrial upgrading and industrial Internet; the digitization of biological world is more and more directed to the cross research of human brain and biology.
In fact, aiot Internet of things is the interface between the digital world and the physical world, and biological computing is the interface between the digital world and the biological world.
And most importantly, its just the beginning. We (China) have been participants in previous industrial revolutions, and this time we have the opportunity to be leaders in some areas.
Qubits: just these three or focus?
Zhang Yaqin: its not so focused any more. If its completely based on personal interests, Ill only do one.
But even so, we are very clear that the core is to provide intermediate and horizontal AI technologies, platforms, modules and algorithms for the three major areas.
Because you are an Industrial Research Institute, you want to use core technology to serve the industry, and make technology become the driving force of industry.
The most basic theory and framework are very important, but they can be done by Tsinghua University, including some departments that have done well now. We dont have to make wheels again.
Qubit: is the refinement of the number of students due to the consideration of tutors energy?
Zhang Yaqin: I think that after too many people, the first one will diverge.
The core research is small and precise at this stage, and does not need to rely on scale effect.
Qubit: will there be a time expectation?
When I started, I had a big goal, but when I started, there was a small goal. But the small goal was not what to accomplish every month and quarter, but what you would certainly go through when you ran a marathon.
A lot of domestic scientific research, used to take the number of papers as the goal, but may forget the starting point for the purpose.
If we have to talk about the goal, I hope there will be 1-2 core original technologies in about 5 years, which will have a great impact on the industry.
Including the transformation of technological achievements, there are Unicorn companies that can be hatched, that is, unicorn companies driven by technology.
In addition, dont be too eager for quick success and instant benefit. Just now, when we say that we only focus on papers and patents, we should not ignore them at all, but we must weigh whether the real value of our work is consistent with our big goal.
My role in this is to constantly remind you to focus on our big goals.
Qubit: does the tutor practice to ensure the direction?
Zhang Yaqin: This is part of the project. In addition, in terms of mechanism, if you refuse to accept short and fast projects, you can take important projects. They can be enterprise, national or social public welfare projects.
Qubit: in the official announcement of Air Research Institute, AI + medical health was a direction. At that time, the epidemic situation had not yet occurred Did the epidemic make you think about it again?
Zhang Yaqin: from the beginning, I was optimistic about AIs transformation of medical care and general health. Even with the current AI technology, many things are helpful.
I saw a good saying before, but now we have done a very accurate weather forecast, mainly based on experience, sensors and big data analysis ability.
This is consistent with the prediction of the epidemic situation. The deeper AI and big data will also have a great impact on molecules, proteins, vaccines and new drugs in the future.
Zhang Yaqin: we can look at things from a larger dimension and see the evolution of the earth, biology and human beings, as well as various social forms.
The most sensible thing is that three things happened a hundred years ago are being repeated in other ways.
Epidemic situation, social fragmentation and disputes, confrontation and conflict between different camps.
There are intuitive and some less intuitive. For example, at the social level, there are data that from March to August, the prices of food and necessities in the United States increased by 2-3 times in less than half a year. And if you look at the stock and market value of technology companies, if you have employees who own stocks, their wealth may increase by 50% So the gap between the rich and the poor is widening.
You know, I have always been optimistic, but now I can only be optimistic in the long term, and I am confused in the short term.
Qubit: is there any change in life?
Therefore, human beings, some countries and nations, actually hope to communicate with each other. I think this is a kind of universal love, and the love of human beings is highlighted.
Qubit: need AI and love?
Zhang Yaqin: Yes, we need AI. We also need to love and be loved.
In the face of major social changes and epidemic situations, in fact, they will return to the most basic human nature. People still need to be together, whether in a region or globally, so I still firmly believe that globalization is the general trend.
Qubit: what suggestions do you have for young people facing graduation from such a severe epidemic?
Zhang Yaqin: its really not easy for this generation. Some have not found a job, some have found a job and lost it, some have found a job and work at home.
Recently, I gave a lecture to a graduate of Columbia Universitys School of engineering. I gave three suggestions on the entropy in the second law of thermodynamics, which represents the degree of chaos in dynamic systems
First, in a world of data explosion and constant change, become a learner with strong adaptability. In a rapidly changing technology industry, most of the knowledge learned five years ago is useless.
The most valuable learning ability is the ability to separate signals from complex noise and the ability to extract entropy from numerous data.
I have an effective habit of spending 10 minutes every morning to find out the latest discoveries or the three most important things for me to learn on the same day.
Second, we should have a unique point of view and perspective. Enter the real world, will naturally start to be polished, to follow the existing trend, into others.
So I want young people to be sharp, angular and different. When I talk to young engineers in person, I look forward to their opinions, their entropy, and their fresh ideas, which are far more important than being smooth, polished and right.
Third, whatever we do, we should uphold morality and humanity.
More than 2000 years ago, the great Greek thinker Socrates regarded morality as the soul of pursuing truth. At about the same time, the great Chinese philosopher Confucius took humanitys benevolence and righteousness as the foundation of social structure.
It is no coincidence that the two great thinkers have the same ideas in different cultures. Today, when we are faced with more choices, confusion and temptation, this becomes more important. Technology is neutral, but innovators have a mission. Technology is a tool, but technicians are for human beings.
Qubit: responding to change with curiosity and learning?
The only thing thats unchangeable in this world is change. Data and information are exploding, and new data generated every year is the sum of data produced in history. Static knowledge learned five years ago is useless, and it is important to learn dynamically.
Education scholars once said, education is what is left after forgetting what we have learned..
In fact, what you learn in college is not important. How to learn is more important. In the era of artificial intelligence, machines are learning, and human beings should learn more.
Qubit: but dont worry?
Its not the battlefield. You have to kill your opponent. Theres no bottom line. Were all losers.
Life is a race field without end, always on the starting line.
Qubit: there is a more specific problem. You have been so busy, but your childrens education has not been delayed. How can you do it?
Zhang Yaqin: Im not so busy, or Im not as busy as you think. I just have more interests and hobbies.
I was very busy for a short time, but in the long run, the schedule was good.
The key is to trust, fully trust, trust the people around you, find someone better than you to do things, and then trust him fully.
The same is true of family education. I fully trust my wife and believe that she makes better decisions than I do. What I do is to be happy with them and encourage them a little more. I seldom criticize children.
Qubit: did the epidemic last a long time?
Zhang Yaqin: This is the first time Ive been with my son for such a long time. Ive been together for six months, every day.
We have defined four joint projects together. The first is barbecue, which has increased by seven or eight kilograms in two weeks. Cooking together is still very interesting. Naturally, there are many topics and languages to communicate with.
The second is to install a computer by himself, which is mainly done by him. The process is very interesting. What kind of template, what kind of power supply and monitor are needed, and how to make the highest performance computer with the lowest price.
Many things have been learned in this process, including what kind of software to use, what kind of configuration is needed, and different memory and graphics cards Later, he developed games and so on.
In addition, he planted vegetables together, but he didnt plant them for a few days. Finally, I was willing to plant them, which was quite good (laughter).
And shooting, boys, dont be too coquettish, experience is still different.
So parents and children, in fact, have a lot of things to do together, so naturally there is a lot to communicate.
Qubit: you dont sound like a tiger dad, either?
Qubit: so now, looking back on the 3.0 you said when you retired, academic, charitable and accompanying your family have made good progress.
Zhang Yaqin: there have been different degrees of progress. Accompanying family members is over fulfilled. For charity, because of the epidemic situation, the first half was played at home, and then the second half was played overseas. It was true that the whole venue was played (laughter), and the amount of money alone was over fulfilled.
Academically, because of the epidemic situation, there was less communication between some places originally planned. However, the core Tsinghua project is still moving forward.
Im in a very happy mood now and theres no pressure.
I have been studying for 20 years before, and I have been working in an enterprise since then. Now I can realize my feelings all the time. I just want to make more contributions.
Source: qubit editor in charge: Mao Xinsi_ NBJS11624