AIR 034 | Interview with Microsoft Research Institute Director Zheng Yu: I prefer to position myself as an innovator

Zheng Yu is the chairman of the program committee of this CCF-GAIR conference and the chief editor of TIST, the top international artificial intelligence journal. Zheng Yu currently focuses on the research and landing of urban computing. This technology can help people greatly improve the quality of urban life through the clever use of the city's massive traffic, air quality, and economic conditions.

In the interview, Zheng Yu said he did not want to be simply labeled as an "entrepreneur" or "academic." For the field of urban computing, he obviously has greater pursuits and expectations.

In the gap between the tight schedule of the GAIR conference, Lei Feng made an exclusive interview with Zheng Yu. This allows us to have a comprehensive understanding of Zheng Yu’s dedicated technology, his perspective, and his expectations for the future.

About city calculation

Lei Fengwang: How did you enter your current research area: urban computing and big data?

Zheng Yu: At the very beginning, there were two reasons: I was very depressed when I was stuck in traffic for a while, and I was a computer student. I was wondering whether I could use my expertise to help the city. Some changes? That would not only be good for everyone, but it would be easier to drive by yourself. At the time, I wondered if I could improve traffic by optimizing some data. Then, just at that time, we got a number of GPS tracks for Beijing taxis, and then we began to try to use them to improve the city's traffic planning. Slowly as more and more of this kind of data we have, we begin to mention environmental protection, people, transportation, and so on.

Lei Feng network: Do you think this kind of technology can bring to people's daily life what substantive improvement?

Zheng Yu: Many, let's say a few of these important areas: Environment everyone can easily think of, use big data to predict haze, predict air quality, to help the government to make decisions: do you want to limit tomorrow? Shouldn't that advise people to reduce outdoor activities? Do you want to remind everyone to close the doors and windows? And where did these pollutants come from?

There is traffic. How do you make the route of the guide vehicle so that everyone can drive faster and more smoothly? How to predict in advance where the crowd will become crowded and avoid stampede? Plan the bus routes rationally and plan the location of each bus and subway platform. These are very obvious aspects.

Lei Fengwang: Do you think that after you have observed and studied so many projects, what conditions do you think a project needs to meet before it can be said to be a worthwhile project?

Zheng Yu: I think there are two aspects. In fact, before MIT had a standard for this issue, the most important one was: “You can tell this to your grandma.” Because if your grandma is on this project If you are interested, it will no doubt indicate that this is a hot issue in society, that is, almost everyone will be interested in it. And the attention of so many people means that its influence will be great.

Second, we need to measure whether it has research value. Although some things are important, they are also very simple to do at the same time and can be easily solved. For example, eating with cutlery is a very important habit. Correspondingly, it is also a very important thing. However, it is very simple. If you are a chopstick, you can do it. It can be said that you don't need to do it. From my perspective, you can do it without using artificial intelligence or big data. Mainly these two considerations.

Lei Feng Network (search "Lei Feng Net" public concern) : Your home page to write these research interests: big data analytics, spatio-temporal data mining, artificial intelligence, and ubiquitous computing, these are more independent research projects or are Serving your main research project: Urban Computing?

Zheng Yu: It's all related to city calculations.

Lei Fengwang: Can you explain how they are related?

Zheng Yu: Actually, it's hard to say so clearly. Big data and artificial intelligence are very broad concepts. There are many intersections but they are not the same. It can be said that machine learning is a branch of artificial intelligence. There are many ways to implement machine learning. Machine learning can be used as a method of data mining, although data mining does not necessarily rely on machine learning. Big data is an overall end-to-end service. These subprojects exist to serve this overall need.

Lei Fengwang: Does the research results of urban computing need to redesign the city according to the results, or does it only need to adjust the existing conditions to some extent?

Zheng Yu: Adjustment is fine. After the model is set, the rest is the adjustment of some parameters, and these models can not only be applied to the current city, but also can be easily adjusted and applied to the new city.

Lei Fengwang: Can you give some examples to illustrate?

Zheng Yu: Two examples can be cited: One is the air quality analysis in Beijing. This is where we aggregate traffic flow data, weather data, and geographic information data to predict changes in air quality over the next 48 hours. This method started with a successful landing in Beijing, and we quickly promoted it to more than 70 cities across the country. This technology is now adopted by the Ministry of Environmental Protection. The second is a big data project in Guiyang. We used a portion of the taxi's GPS trajectory as input to calculate the speed, flow, and exhaust emissions for each street in the entire city.

From academic to industry

Lei Fengwang: At present, what stage do you think this project is going through?

Zheng Yu: Basically, the air quality forecast has been completed on the first phase of our air quality forecast and we are preparing to do the second phase. In the first phase, air quality analysis and forecast of 70 cities were done. In the second phase, we have to do 300 cities. Our first phase in Guiyang is basically over. A Guiyang urban big data platform was built to manage these data.

Lei Fengwang: In addition to air quality and driving data, what data do you intend to collect in the future?

Zheng Yu: In fact, these data do not require us to actively install any equipment to collect. The existing equipment is enough to collect them. They are already there. We just use it. In the future, we will apply to planning, environmental protection, energy consumption, economics, education, and the analysis and application of these data. We are now doing a dynamic carpooling service. In fact, this is also a way to increase traffic and reduce congestion. With respect to energy consumption, we are now doing automotive traffic analysis. In economic terms, what we are currently doing is an estimate of the value of the property.

Lei Fengwang: Why did you choose Microsoft Research as your place for research?

Zheng Yu: The first of the benefits of Microsoft Research is that it can connect internationally and domestically. It has both an international perspective and a strong bond with the domestic market. The second is that it can span industrial and academic circles. Microsoft Research has the ability to deploy technology to the market and industrialize it. So this is a good opportunity. In addition, there are still many excellent talents in Microsoft. In an environment, I think people are the most important. If the people around you are smart, you will get a lot of help.

Lei Feng network: In contrast, the same academic-based Chinese Academy of Sciences compared with Microsoft Research, which do you think is more suitable for you?

Zheng Yu: I may think that Microsoft will be more appropriate. Because relatively speaking, the Chinese Academy of Sciences still focuses on research and exploration, rather than industrial layout. When a lot of engineering power is needed, the Chinese Academy of Sciences is not necessarily the best place to do it. It also involves the market. For example, you can't do anything on your own. When we do things to a certain degree, we need to deal with the government. We need our R&D, sales, marketing, and government to interact together. At this time, we need specialized people to do special things. We, these people come to do research and development, and let us those responsible for the market to communicate with the government and negotiate prices.

Academic, research, self

Lei Fengwang: How do you think China's current academic atmosphere is compared with foreign countries?

Zheng Yu: The current academic atmosphere in China has improved a lot. China’s research funding is actually more abundant than foreign professors. The key is that China has a lot of data and it is much easier to get data than abroad. And Chinese students are not themselves worse than foreign countries. In fact, many outstanding students from abroad are also Chinese in the past. So China's environment is now getting better, people are staying, data is there, money is there, and the level of teachers is also much higher. It can be said that the conditions have been greatly improved. I think it is a good opportunity for China.

Lei Feng Network: Is there still some gap?

Zheng Yu: I think the main reason for the disparity is still the strength of the teachers in the school. Although the level of college teachers has improved a lot, there are still gaps compared with the top level in foreign countries and the problems of the policy environment of some universities.

Foreign policies and financial systems will be more liberal. It is that some of its projects allow some compensation for the people involved in the project. However, some systems in China may stipulate that you are not allowed to pay the participating personnel. Sometimes people who are really involved may not get financial returns, including teachers. Funds are only allowed to invest in the purchase of equipment. In fact, this is not a particularly big problem, but to some extent, it may reduce some of the enthusiasm.

Lei Feng: Do you position yourself more as a scholar or someone who wants to advance into the industry?

Zheng Yu: I prefer to position myself as an "innovator." Being an innovator requires several skills: First, you need solid academic knowledge. There must be technical skills. Second, if you really want to complete your innovation, you must really deploy your product. It cannot be said that only one paper was published or it was finished. If you don't see anything written, it's not innovation. At this time, you need practical skills. Third, when you are an innovator to promote your innovation, you are actually a salesperson and you need to communicate and collaborate with others. So this is a manifestation of comprehensive capabilities. I hope more that I am a person with comprehensive ability, not that I am an expert in the field of machine learning.

Lei Fengwang: As an innovator, what are your requirements and expectations for yourself?

Zheng Yu: My greatest hope and goal is to lay a solid foundation for urban computing and lay a solid foundation. In fact, this area has actually done relatively well and is relatively prosperous. There have even been professional directions for city computing abroad. Now I am writing a book called "Urban Computing," the first book in the world about this area. This book will serve as a teaching material for colleges and universities, so I hope that through this, we can build this field so that more people can do what they want to do.

Low Voltage Drives

The low voltage inverter produced by FGI covers every automation application requirement in the industrial plant. This low voltage converter is mainly used for controlling and adjusting the speed and torque of three-phase AC asynchronous motors. Using the latest technological advancements in AC motor speed control, the FGI high performance low voltage drives meet the needs of smart machines by increasing system reliability, improving process control and minimizing total installed project costs.

Low Voltage Drives,Low Voltage Vfd,Low Voltage Ac Drive,Low Voltage Inverters,Low Voltage Converter,Low Voltage Boost Converter

FGI SCIENCE AND TECHNOLOGY CO., LTD , https://www.fgi-tech.com