Baidu Duan Runyao: 4 factors make the emergence of quantum computing inevitable | MEET2023

Fengse compiled from MEET2023
qubits | Public account QbitAI

From the popular topic of "quantum supremacy" to this year's Nobel Prize in Physics going to the field of quantum information science.

Quantum computing , a technical term, has long become one of the most talked about scientific and technological issues at the moment.

A series of latest technological advances have also given rise to a new problem:

How far is quantum computing from our lives?

In fact, some data show that by around 2031, a market size of nearly 800 billion will be directly related to quantum computing.

This means that this field is about to usher in a grand journey of implementation.

So, how do we take the first step and build a true quantum computer?

Duan Runyao, director of the Baidu Institute of Quantum Computing, studied under Professor Ying Mingsheng. He is a tenured professor at the University of Technology, Sydney, the founding director of the Quantum Software and Information Center, and a Future Fellow of the Australian Asian Research Council.

He has been engaged in the fields of quantum computing and quantum information technology since 2001 and has made original contributions; Baidu, where he works, has taken the lead in providing a complete set of industrial solutions for quantum computing from underlying hardware to upper-layer applications.

So at the MEET 2023 Smart Future Conference , Duan Runyao shared some experiences and opinions using Baidu as an example. He said:

The last few years have been very critical years for quantum to truly begin to be integrated with computing. The emergence of quantum computing is inevitable in many aspects.

If we want to continue the advantages of Moore's Law and cheap computing power, we need to consider a new computing model, which is quantum computing.

And if he wants to build a real quantum computer, his conclusion is simple:

It's a smart brain, matched with sufficient resources and the best available technology.

In order to fully reflect Duan Runyao's thinking on solutions for the industrialization of quantum computing from hardware to software and applications , Qubits has edited and organized the content of his speech without changing the original meaning.

About MEET Intelligent Future Conference: MEET Conference is a top business summit in the field of intelligent technology hosted by Qubit, dedicated to exploring the implementation and industry applications of cutting-edge technology. This year, dozens of mainstream media and live broadcast platforms reported and broadcast the MEET2023 conference live, attracting more than 3 million industry users to participate online, and the total exposure of the entire network exceeded 20 million.

Key points

• To continue the advantages of Moore's Law and cheap computing power, we need to consider a new computing model, namely quantum computing.

• We are already at a very critical point where the advantages of quantum computing begin to overtake classical computing. If we break through this point, there will be many possibilities in the future.

• By around 2031, nearly 800 billion in market size will be directly related to quantum computing; nearly 70% of companies want to make some arrangements in this area.

• Quantum computing and artificial intelligence are in a mutually reinforcing and entangled relationship. They can be connected from the upper application, the entire platform framework in the middle, to the lower hardware level.

• There is a lot of overall investment in quantum computing, but investment in specific algorithms and software is limited. This aspect needs to be strengthened and has reached a critical stage.

• How to really build a quantum computer? My conclusion is actually very simple, it is a smart mind, coupled with sufficient resources and the best available technology.

(The following is the full text of Director Duan Runyao’s speech)

Why is the emergence of quantum computing inevitable?

When I saw this icon (conference icon) today , it reminded me of quantum entanglement in quantum information.

In fact, it can be said that the quantum computing we are discussing today itself is the entanglement of computing and quantum information .

On the one hand, from the knotted ropes used for counting in ancient times, to the abacus and slide rules later, to today's computers and supercomputers, computing has covered every aspect of our society.

On the other hand, quantum is actually our new understanding of the world.

We know that the primitive era only had a superficial understanding of the world through experience. It was only later that Newtonian mechanics and Maxwell’s electromagnetic theory began to appear. In the last century, we were finally able to discover that the world actually follows the laws of quantum mechanics.

Since the emergence of quantum mechanics in the 1920s, it has actually had an earth-shaking impact on our world. This is the so-called first quantum revolution .

Things like lasers, transistors, nuclear magnetic resonance, and atomic energy that we are familiar with today are all based on the direct use of quantum mechanics, especially the Internet used today.

This first quantum revolution has basically changed every aspect of our entire society.

Now, our era has actually entered the stage of the second quantum revolution , or the new quantum revolution.

At this stage, we are no longer satisfied with making quantum mechanical predictions on a piece of material as a whole. We hope to directly control a specific system and use such a system for processing. In the process, some quantum properties like quantum entanglement really make a difference.

The past few years should be very critical for the real beginning of the integration of quantum and computing, especially the well-known quantum computer "Platanus" developed by Google, and the "Zu Chongzhihao" developed by the University of Science and Technology of China. These all prove that when quantum and computing are combined When together, superpowers can truly emerge on certain issues.

After the true combination of quantum and computing, new disciplines will emerge. We also believe that this thing will bring us a whole new world because it can bring exponential improvements in computing power.

After talking so much, you may have a general impression that the combination of quantum and computing is very important, but how important is it?

In fact, the emergence of quantum computing is essential.

On the one hand, when the size of our chip reaches a certain level, such as 1 nanometer, it enters the scale of quantum mechanics. In this scale, we need to consider using quantum mechanics to design the chip, instead of just relying on The traditional way.

So at this time, we are naturally required. If we want to continue the advantages of Moore's Law and continue the cheap computing power, we need to consider a new computing model, which is quantum computing.

Of course, this is also related to energy consumption. Because everyone knows that computing consumes a lot of energy, but if quantum computing is used, it is a reversible computing mode, and the final energy consumption can be very small.

On the other hand, our data volume is already very large. As mentioned in the reports just now, various large models require a lot of calculations.

In fact, when we want to simulate a quantum system, even a very small system, such as 300 qubits (qubits are the quantum counterparts of classical bits) , the storage required to express the state of such a system The amount already exceeds the number of atoms visible in our entire universe.

In addition, quantum has many positive effects. For example, there are many algorithms that we can theoretically prove its advantages, especially its exponential advantages on some specific problems.

And we use the Internet every day to conduct various transactions or transmit various information. In this process, the security of information is very important.

But unfortunately, quantum computers can actually break the financial RSA system.

For example, if we do the prime number decomposition of a 1,000-digit large number, it may take a long time using a traditional computer, which is longer than the length of the universe. But if a medium-scale quantum computer could be built, it would probably only take a few seconds or even a few seconds to solve this problem.

By 2031, the quantum computing market will reach 800 billion

In fact, we are at a very critical point today, where the advantages of quantum computing begin to surpass classical computing. If we break through this point, there will be many possibilities in the future.

It is precisely because of this possibility that quantum technology has always attracted great attention from the current industry.

As you can see, in the "14th Five-Year Plan" and the 2035 long-term goals, quantum information, integrated circuits, and artificial intelligence are particularly listed as one of the three priority development areas.

Major foreign countries such as Europe and the United States have also invested heavily in this aspect.

Everyone in the industry has realized that it is likely to bring about a new information revolution. Large domestic and foreign companies have already made plans in this area, including Baidu , which also began to truly enter this field in 2018.

In addition, there are many very good start-ups also doing some research.

Here are two pieces of data. After synthesizing various data, we found:

By around 2031, the market size of nearly 800 billion will be directly related to quantum computing; nearly 70% of companies want to make some arrangements in this area.

So it's a very exciting thing. So in what specific areas can it be used?

We know that quantum computing will bring about changes in everything from hardware to software and algorithms, so basically we can say that many important fields are related to it.

Especially the several directions I have briefly listed here, such as chemical industry and medicine, especially drug research and development issues; or some optimization issues in financial technology; or including chemical simulation issues in new energy batteries, and information security... these All aspects can be closely related to quantum computing.

In addition, in fact, quantum computing and artificial intelligence are also in a mutually reinforcing and entangled relationship. Not only the above applications, but also the entire platform framework in the middle and the underlying hardware level can have mutual connections. This is a quantum pair. The impact of our artificial intelligence.

On the other hand, quantum can also benefit from some of our existing very advanced artificial intelligence technologies. In particular, deep learning can help us study many properties of quantum systems, including quantum algorithms and quantum protocols.

For example, Baidu has built a quantum machine learning platform called Liangpiao in 2020 .

The industrialization of quantum computing needs to solve three problems

We currently have so many opportunities to rely on, and we also know that quantum is about to enter our real life, but how many students here or online have actually used quantum computers?

I believe the number of people is still very limited.

Many times we may only see relevant reports from the news or newspapers, but if we really want to industrialize it, we need to bring quantum computers from the laboratory into our actual lives.

But what is the problem in the laboratory? The platform is single and unstable... There are many problems that make it difficult to output quantum computing power.

What needs to be solved on our industrialization path? A series of problems need to be solved from hardware to software and applications.

For example, in terms of hardware, we need to solve the problem of stability, that is, whether your device can continuously output quantum computing power. If it is turned on today and down again tomorrow, this will not work.

Second, you also need to provide a platform, a platform for hardware. In addition, you also need to ensure that this chip can have an automated design method that can continuously provide you with new generations. This is very important.

At the software level, a particularly important point is whether we can have a good software platform that allows our programmers to move from classical programming to quantum programming. In addition, it also includes how to make software interfaces, how to make error corrections, etc.

At the application level, a good algorithm library is needed to provide truly advantageous scenarios.

In short, the industrialization scenario poses multi-dimensional challenges in scientific research, technology and engineering.

Baidu proposed the QIAN strategy . Q is quantum algorithm, quantum AI and quantum architecture. I is the software and hardware infrastructure. In addition, some scenarios with significant application value must be truly identified. A refers to some important actual applications. Finally, of course, there is the network. N refers to the quantum Internet. In addition, we also need to establish an ecological network.

Among them, you can see three interpretations of "QIAN": on the one hand, it conforms to the general trend of technological development, which is the "qian" of "Qiankun"; on the other hand, it is the "front" of "forward-looking", truly focusing on forward-looking applications; The last aspect is that the "money" of "money" can truly make good use of quantum ecological resources.

After more than 4 years of development, the entire Baidu Quantum has been fully integrated from the above applications to the underlying hardware. It includes not only software, but also hardware, as well as Quang Yi Jian tutorials, quantum operating system platforms, tool sets, etc.

That is to say, under the guidance of such a platform, we finally determined to unify this whole set of things, so we hope to truly integrate software, hardware, and applications and make them available to everyone, truly realizing the integration of software and hardware. Put quantum computing power into the hands of every user.

We launched a highly integrated industrial-grade quantum computer called "Qianshi" at the first quantum developer conference in August this year .

Why is it said to be an industrial-grade quantum computer? In fact, it is because it can provide services continuously and stably. Although the current chip capabilities are still relatively limited, we have been able to continuously and stably provide services to users.

In fact, as long as everyone is interested in quantum information, they can open the mobile phone and download the Liangyifu APP, and you can quickly learn some of the most basic quantum knowledge.

We have provided 5 interesting applications, from quantum superposition to quantum entanglement, to quantum search, and there will be a series of quantum neural networks for everyone to use. We believe that through such efforts, quantum computing will be within reach.

So companies like Baidu have begun to work on quantum industrialization, but more companies themselves may also hope to truly introduce quantum things into their own company services. What should we do about this? In other words, how can we help other industries achieve industrial quantumization?

I mentioned this data just now. Nearly 70% of enterprises have this demand. At this time, everyone will find that there is a big gap that needs to be crossed.

In August this year, we released the world's first integrated quantum software and hardware solution called Liangxi , which can very conveniently realize the enterprise's quantum layout.

This platform now has great advantages. First of all, it is a full platform. It can be adapted to all kinds of hardware, whether it is mobile phones or PCs.

Secondly, it actually has a complete set of software and hardware technologies.

On the other hand, it can also truly provide enterprises with practical services, such as privatized customized cloud services, which can help them with overall deployment and one-stop upgrades.

Specifically, there are three steps. First, this company can come and discuss with us what kind of impact quantum computing will have on your industry and your company, and we can formulate a strategy; second, it has been confirmed to be a trend, so we will You can consider truly identifying some core problems in the enterprise and using our quantum methods to develop solutions for you; the third step is specific implementation.

We hope that through technologies like our Baidu Quantum Platform, including Qianshi’s actual quantum products, we can eventually produce solutions like Quantum Xi’s, which will help us in all walks of life.

"A smart brain + enough resources + the best available technology = a quantum computer"

At the end of this report, I would like to make some outlook.

In fact, I have been thinking about a question for more than 20 years: How can we really build a quantum computer?

My conclusion is actually very simple, that is, a smart brain, coupled with sufficient resources and the best existing technology, can make a quantum computer.

In this sense, there are many different combinations around the world that can make such a machine, and the process is equivalent.

What are the current problems with quantum computers?

First of all, talent is very scarce. If you study quantum computing, there are several offers waiting for you on the market.

The second is funding. The overall investment in quantum computing is huge, but investment in specific algorithms and software is limited, and this aspect needs to be strengthened.

In fact, a lot of capital can consider quantum computing, because it may have been a bit early before, but now it has reached a critical stage.

In addition, we actually hope to build an open and sustainable quantum computing ecosystem through our Baidu Quantum Platform.

In this way, we not only provide services to our developers and our users, we also hope to establish cooperative relationships with academia, our senior experts (including our consultant Artur Ekert) , and experts in quantum cryptography.

In addition, we also need cooperation with the hardware team, because we mainly focus on software. Although we also have our own hardware, we hope to have more in-depth cooperation with the hardware team to make quantum chips better.

Of course, another aspect is that cooperation with universities will always be the source of our innovation.

Thank you again for listening, thank you!

click here