Forefront | SenseTime CEO Xu Li: Advocating that the industry and society should uphold the "development" artificial intelligence governance concept

Forefront | SenseTime CEO Xu Li: Advocating that the industry and society should uphold the “development” artificial intelligence governance concept

#Forefront #SenseTime #CEO #Advocating #industry #society #uphold #development #artificial #intelligence #governance #concept

On July 8, the three-day 2021 World Artificial Intelligence Conference (WAIC) kicked off at the Shanghai World Expo Pavilion.

Last year, WAIC used online as its main position. After returning to offline exhibitions this year, WAIC’s guest lineup is still not inferior to the past. The guests who delivered speeches on the opening ceremony included Baidu founder Robin Li, Gree Electric Chairman and President Dong Mingzhu, Huawei Rotating Chairman Hu Houkun, and SenseTime’s co-founder and CEO Xu Li and other industry celebrities.

In addition, Tencent Chairman and CEO Ma Huateng, Sequoia Capital Global Managing Partner Shen Nanpeng, and Software Banking Group Representative President Sun Zhengyi participated in the opening ceremony agenda in the form of video/connection.

SenseTime’s co-founder and CEO Xu Li delivered a keynote speech titled “A New Paradigm of Artificial Intelligence Innovation-Machine Conjecture”. He said: Artificial intelligence breakthroughs will be able to regularly expand the boundaries of human cognition. He pointed out that whether it is the induction of past experience, deduction of reasoning, or the use of big data and simulations, the four scientific research innovation paradigms in the usual sense have established models, but there are many disruptive scientific innovations and major breakthroughs in human history. , Are all counter-consensus and unpredictable, mostly derived from “genius-like conjectures”, just like the apple smashed on Newton’s head and the light that Einstein dreamed of riding in his childhood.

Xu Li, Co-founder and CEO of SenseTime

“The reason for this phenomenon lies in the limitation of human understanding of the unknown world. But in today’s artificial intelligence era, machines can also make guesses, which is expected to help us discover the essence of scientific laws earlier and explore and discover faster Unknown, this is the new paradigm of innovation brought about by artificial intelligence.” Xu Li said.

With the development of general artificial intelligence technology, AI algorithms are no longer entirely dependent on labeled big data, but the exploration of huge unknown possibilities must rely on large computing power. Xu Li said that the SenseCore AI large device created by SenseTime provides a powerful boost to “machine conjecture”. Basic disciplines such as life sciences, material sciences, and medicine will be the first to benefit from “machine conjecture” and enter a new paradigm of scientific innovation.

Take the development of autonomous driving as an example. From unmanned parking, to unmanned connection in closed parks, to autonomous driving on open roads, it is a process of applying data to define the boundaries and specifications of the technology, so as to gradually achieve The controllable development of technology can better promote the progress of the industry.

The innovation of artificial intelligence comes from the subversion of past cognition, so the issue of ethical governance in industry applications is becoming more and more important. “People-oriented, technology-controllable, and sustainable” is an internationally agreed framework for technological ethics governance, but at different stages of development, balanced development of the three is required.

In this regard, Xu Li advocates that the industry and society should uphold the “development” ethical governance concept of artificial intelligence. One is that when talking about ethical governance, we must consider the goal of inclusive development, that is, use artificial intelligence technology to promote social progress; the second is While implementing advanced technologies, it is necessary to consider the rapidity of industry changes and find out the balance of development under different governance frameworks for different development stages.

Xu Li finally stated that the “machine’s conjecture” played the role of an apple hit on Newton’s head, and Shanghai’s tolerant, open and innovative temperament gave this apple the best soil to help it take root and grow out. Apple tree promotes innovation in various industries and develops inclusive and sustainable artificial intelligence.

This year’s WAIC exhibition set up a “1+6” section, and “AI@上海” themed exhibition area, innovative technology, new kinetic energy, smart economy, new vitality, new picture of a better life, new development of coordinated governance, building a new ecological future, and a new harmonious humanity. There are 6 sections in the chapter.

Among the exhibitors, the number of offline exhibitors is expected to exceed 300 this year. Among them, the proportion of enterprises participating for the first time is more than 40%, and the proportion of enterprises outside Shanghai and foreign enterprises is more than 50%.

The following is the full text of SenseTime’s co-founder and CEO Xu Li’s speech at WAIC 2021:

Dear Secretary Li Qiang, Minister Xiao, Mayor Gong Zheng, leaders and guests, good morning.

I am very fortunate to stand on this podium and share our thoughts on the paradigm of artificial intelligence innovation. SenseTime is very fortunate to witness the development of WAIC in the past 4 years with everyone. I looked through my photo album and found that I took photos every year, and then the theme of the conference was taken in the photos every year. From 2018 to artificial intelligence empowering a new era, to 2019 we are discussing the infinite possibilities of artificial intelligence, and then to 2020 and 2021 to talk about homes and cities, artificial intelligence seems to be getting closer to life.

In fact, artificial intelligence conferences are held all over the world, and there are many places across the country, but if we look at the popularity of search engines, we will find that all the peaks are during our WAIC conference, which means that WAIC has already Invisibly become the most important participant, witness, and recorder in the history of artificial intelligence.

Then talk about innovation. In fact, I think artificial intelligence is an innovation park. We have been talking about the paradigm of innovation in the past, from inductive summation to deductive reasoning. When computers are available, deductive reasoning can be simulated by computers. In other words, computers can be used to make big data-driven algorithms for induction and summarization. This is our Innovation and creation are generally referred to as four ways of innovation. But the real subversive innovation does not come from here. It comes from the brains of geniuses, or from the incredible conjectures of geniuses, from “thought experiments”.

So today I want to discuss this matter, that is, with artificial intelligence and supercomputers, can we re-ask the question asked by Alan Turing in 1950, “Can machines think”, then our proposition is simpler Some, can you ask “Will the machine guess?” I think the answer is yes, because in some fields, machines have given us humans a very good model, making it reverse the progress of our human sciences.

What are the necessary conditions for machine guessing?

First of all, with the development of computing power in the past 20 years, we can see that the demand for computing power for the best artificial intelligence algorithms has increased by nearly 1 million times in the past 10 years. This is a huge breakthrough. It stands to reason that the more advanced the algorithm, the less computing power will be required, but on the contrary, the larger the unknown space we explore, the greater the computing power is needed, which is irrelevant to the data. Because many algorithms already rely on small data, or even do not require human data, but this kind of exploration actually gives us a possibility to update iterative cognition.

Therefore, I said that artificial intelligence’s computing power is a basic element to promote conjecture. Why is it called a big device? I want to make an analogy to a particle collider, which is a random possibility, creating new particles to explore the unknown world.

Second, since we are all here to guess, many things are not certain, so when defining the boundaries of industrial applications in a true sense, or when defining it as unreliable, we need to put them in the industry. In fact, there are many applications in our human history to say that first there are conjecture applications, and finally we can give you a proper explanation.

Take airplanes, for example. When the Wright brothers invented airplanes, they certainly didn’t know the principle. Even to this day, we can’t use fluid mechanics to explain the dynamics of airplane take-off, but this does not prevent airplane manufacturing companies from making safe and controllable airplanes. In fact, among the many artificial intelligence applications we are familiar with, most of them are based on such conjectures.

Let’s take a look at a few simple examples. A very small scene in autonomous driving is called automatic parking. Old drivers can often sum up the rules of parking, which makes sense. However, these laws cannot be transformed into computers in time. Computers are a set of laws evolved by themselves. Even our scientists have no way to truly explain the logic of computer reversing, but this does not prevent our automatic parking. The scene has fallen into place in many places.

Extending to the unmanned driving we are talking about now, I saw that there are many unmanned cars off the field, and our own unmanned AR minibus also started to ride yesterday, and to show our augmented reality. The first thing the passengers ask is, how many kilometers can your car travel without someone taking care of it? I feel very relieved. This is precisely because we are cautiously gradually opening up the application scenarios, so our understanding of the unmanned driving scenario has spread to the general public, and in the future, we will embrace such a scenario iteration more so that we can truly explore it. The possible boundaries of technology.

Another more complex application is decision-making in an open environment. For example, in our game, when there is still someone to guide the reversing, the space complexity of the game is much higher than that of reversing, or even Go. . In this process, it is actually difficult for humans to give a standard answer, which is like the application of intelligent transportation that we will talk about later.

Under normal circumstances, when we solve the problem of traffic signal control, it has been difficult for humans to give a perfect answer. We are trying to use artificial intelligence conjectures to solve this problem in a small area, which can save the local average waiting time by half, which is close to 20 minutes. So can it be extended to a larger area? We still wait for us to do it. More temptation. But the machine’s conjecture in this direction has given us an unexpected surprise.

The other type of application, I believe everyone may think it is a big data-driven application, including the two-network construction of a smart city in Shanghai, including our smart community. However, when we actually entered these applications, we found that it is not that big data. What we have to solve is the daily long-tail low-frequency, just-needed applications.

For example, when we solve objects that fall from a high altitude, it is impossible to have too much data on such fallen objects; when we solve fires and we solve the community care problems of elderly people falling, there are often zero data and small data. , Then the machine can only make extensions and generalizations through general technology to make possible guesses about these scenarios.

Since machines are in many cases, we currently cannot have a complete control and complete iteration, so SenseTime also upholds the key to promoting better ethical development of artificial intelligence, and launched our controllable, green and environmentally friendly , Sustainable, people-oriented ethical development of artificial intelligence.

What I want to emphasize here is that we should look at the ethical governance of artificial intelligence from a development perspective. Development has two meanings. First, we must take development as the goal. All ethical governance is to make artificial intelligence better serve our society and better promote social development. If you don’t talk about goals, it’s actually very easy to meet any governance condition, but a single optimization will often cause everyone to get into trouble.

Second, the development of artificial intelligence is rapid, so we have to balance various governance frameworks with a development perspective, and choose different governance policies at different stages. Just as Secretary Li Qiang said just now, we need to promote agile governance , From completing such a task.

A negative example is when the first car was built in Britain in 1865, the United Kingdom introduced the red flag law. There was a person waving a red flag 50 yards in front of the car, so the speed of the car could not be faster than that of a person. Of course we have achieved complete safety and control, but the UK has also lost the bonus time for automobile development.

So today, I think that artificial intelligence and machine guesses may be the apple that hit Newton’s head. Shanghai’s tolerance and openness are its best soil, enabling it to find more innovative results. We may use the innovation of artificial intelligence to promote the inclusiveness of artificial intelligence, so that artificial intelligence has more influence on many industries.

thank you all.

The speech on Pratt & Whitney Artificial Intelligence is over, but I believe the discussion on Pratt & Whitney Artificial Intelligence has just begun. Then we will invite an important guest, Sequoia Capital Global Managing Partner, Sequoia Capital China Fund Mr. Shen Nanpeng, the managing partner.

.


Related Posts