Read in one article: Ten questions about quantum computing

Jointly produced by Tencent Research Institute and Tencent Quantum Lab

With the in-depth development of artificial intelligence large models, quantum computing, brain-like intelligence, cloud native, digital engines, audio and video and other technologies, new technologies, new models and new formats continue to emerge. In view of the field of Shangda, we have planned a popular science series column "T-chat Frontier Hot Questions and Answers", in the form of ten questions and ten answers, in plain language, to communicate with you the latest and hottest cutting-edge science and technology topics. This is the second in the T-Chat series. You are also welcome to guide, and put forward questions of interest to study together, continue to send generations of cognition, and make progress together.

In addition to the ingenious ChatGPT, another hot topic that has attracted attention recently is quantum computers. While playing the future computer "MOSS" with great ability and intelligence in the science fiction movie "The Wandering Earth 2", it also appeared on the cover of the famous journal Time magazine in February with the title "The Quantum Leap". But for the general public, the interpretation and interpretation seen on the Internet can only bring an unfathomable sense of ambiguity; A series of concepts that seem to understand and not understand can only be reduced to a single word in the conversation that seems to understand and does not understand.

So, how exactly should we understand quantum computers? What does it have to do with ChatGPT? In the new stage of the digital age, how can we correctly understand it, use it, and have a more realistic expectation of it? Quantum computing ten questions and answers for you.

Article Table of Contents:

1. What exactly are "quantum" and "quantum computing"?

2. Why are quantum computers important to us?

3. What is the difference between quantum computers and our common computers?

4. The larger the "quantum volume", the more powerful the quantum computer?

5. Are quantum computers really "omnipotent"?

6. How to create an "all-knowing and all-powerful" "computer of the future"?

7. Is quantum computing now in practice?

8. Is quantum computing more suitable for simulation?

9. What companies and institutions at home and abroad are currently doing quantum computing?

10. What remains to be done before quantum computing can be put into practice? How can we accelerate real-world adoption?

"Quantum" and "quantum computing",

What is it?

Quantum computing is a new computing mode based on the principles of quantum mechanics that controls a certain number of quantum units. Quantum computers are computer systems used to implement and use quantum computing power.

Quantum is an important concept in modern physics, considered to be the smallest and indivisible unit of physical quantity that makes up the vast universe and everything in the world. For example, sunlight is composed of photons, and the energy carried by photons cannot be further divided at present, so photons can be called quantum. In addition, electrons, neutrinos, and quarks can all be called quantum. Microscopic particles such as atoms, ions, protons, and neutrons carry different energies due to different conditions, so it is not easy to judge whether they can be called quantum.

The physical properties of quantum are completely different from objects in nature (such as footballs, beverage bottles), and even beyond our common sense and imagination: for example, quantum can fly in the air like a kicked football, or fly forward fluctuating up and down like water waves; For example, if a beverage bottle is placed on a horizontal table, the bottle can only maintain one of the three states of "upright", "upside-down" and "lying horizontally", while the quantum that also stands on the table can maintain the three states of "upright", "upside-down" and "lying horizontally" at the same time.

More than 100 years ago, Planck, Einstein and other famous physicists discovered the unique physical properties of these microscopic particles, and finally summarized these properties into "quantum mechanics", one of the two cornerstones of modern physics. We can refer to the technology based on the principles of quantum mechanics for information application, collectively referred to as quantum information technology, mainly including quantum computing, quantum communication and quantum measurement.

Quantum computers why

Important to us?

From a scientific and technological point of view, mastering quantum computing technology will grasp the opportunity for future development. In 2023, we are already in the digital world, where cities, industrial production, and people's lives are constantly generating, processing and exchanging information. The computing power (referred to as "computing power") widely present in smart devices, networks, servers and cloud platforms is the most fundamental and critical capability to support all this.

At present, computing power is based on the well-known computing technology with CPU as the core; In the future, quantum computing technology is a more advanced and promising next-generation computing technology that will realize future computing power. Quantum computing technology is like a shining pearl at the pinnacle of science and technology, attracting the attention of academia and industry.

From a societal perspective, quantum computers are representative of advanced digital productivity. Some quantum computer prototypes announced by the industry, such as Google's "Hanging Suzuki", IBM Q System and the mainland Chinese Academy of Sciences' "Chapter Nine", have gradually demonstrated some computing power advantages in scientific research experiments: when completing specific computing tasks, they can show thousands of times higher computing efficiency than supercomputers. Materials, medicine, finance, logistics, manufacturing and many other industries have begun to pay attention, explore and try. The influence of quantum computers in the fields of information security and high-performance computing has basically become the consensus of the industry and has attracted great attention from all countries. At present, more than ten countries have announced strategies related to quantum computing, and the research and development of quantum computers has become an international hot spot for a while.

For our people, quantum computers are still a little far away, but they are worth looking forward to. Although the mature application of quantum computers is expected to take decades or even longer, its gradual application in some subdivided scenarios should bring us some surprises in the near future. For example, assist in the development of special drugs for incurable diseases, accelerate the development of new fabrics that are lighter and more comfortable, and so on. Of course, quantum computers as a commodity into thousands of households, may still be more distant things, but no one can predict the speed of technological development and popularization.

Quantum computers and our common ones

Computers, what are the differences?

The smallest computing elements of quantum computers are much more advanced and powerful than ordinary computers (smartphones, laptops, industrial computers, servers, etc.). This is the most fundamental difference between the two, and the other 7 differences are derived from this.

The answer to this question is a little tricky, but we try to explain it in a layman's way. If you can patiently read the following narrative and try to understand it, then you will initially grasp the code to understand and even predict the future trend of digital civilization.

First, to understand the difference between the two, you need to understand some basics:

Mathematics, the oldest science of mankind, is also one of the most basic disciplines of human science, helping people understand the world from the perspective of numbers and survive in the world.

Computational tools are tools invented by people in order to make better use of mathematics. Such as abacus, electronic calculators, computers. Usually, people first disassemble and abstract the problem to be solved into several input conditions and mathematical models, and then use calculation tools to calculate, and finally obtain a reference for solving the problem from the calculation results.

The smallest computing element is the smallest physical element that can be manipulated (written, stored, read) by the calculation tool and represents the calculated value. The advanced nature of the minimum computing element largely determines the upper limit of the computing power of a class of computing tools. So what is the smallest calculation unit for different calculation tools?

Abacus: The smallest calculating element of the abacus is an abacus bead in the lower row of the abacus, indicating the calculation value is 1, dialing up means +1, dialing down means -1;

Figure: Abacus on an abacus

Ordinary computer: For ordinary computers such as smartphones, laptops, industrial computers, and servers that we commonly use now, the smallest computing unit is a transistor on the central processing unit (CPU), which can represent a value of 0 or 1. When the transistor is energized, it means 1, and when the transistor is powered off, it means 0. A transistor, in classical computing principles, represents a binary digit called a bit (English: Bit);

Quantum computer: The smallest computing unit of a quantum computer is a quantum on a quantum processor (QPU), the core component of a quantum computer, which can represent a value of 0, or 1, or a combination of 0 and 1 in a series of probabilities (e.g., 80% 0+20% 1, or 40% 0+60% 1, etc.). It's more difficult to understand here, you don't need to understand, just know). When people use the input device of a quantum computer to emit specific electrical pulses to quanta, they can make the quantum represent different values. A quantum, in the principles of quantum computing, represents a smallest unit of measurement, called a qubit.

So far, we can find that a quantum, which can represent a lot more computational values, than a transistor. This is like Sun Wukong, compared with ordinary people, the ability is greater; Further, because of the "superposition state" nature of quantums, a quantum can maintain multiple states at the same time, which means that it can represent multiple computational values at the same time. It's like Sun Wukong not only has great skills, but also has multiple doppelgangers at the same time, each of which is as powerful as him.

Figure: The "superposition state" of quantum is like the multiple doppelgangers of the Monkey King

After understanding the fundamental differences, we begin to reveal the rest of the answers to "what are the differences between quantum computers and our common computers":

1. Different technologies for manufacturing the smallest computing elements (quantum chip process): the chips of quantum computers are currently prepared by various technical routes such as superconductivity, optical quantum, and ion traps, while the chips of classical computers mainly use lithography machines;

2. The minimum number of computing elements contained in a single chip is different: the latest NVIDIA GPU H100 contains about 50 billion transistors, while IBM's latest quantum chip Osprey only contains 433 quantums;

3. Different computing power: Quantum computers with only about 50 quanta can already be thousands or even tens of thousands of times faster than the most powerful classical computers when completing specific computing tasks;

4. Different hardware systems: Due to different chip processes and early research and development, compared with classical computers, current quantum computers require special supporting equipment and systems. For example: chiller, digital-analog control system and so on. This also makes the current quantum computer appear bulkier in appearance, some like the electronic computer that came out 60 years ago;

Pictured: An electronic computer more than 60 years ago

5. Different software systems: On the one hand, quantum computing algorithms and applications need to adapt to the principle of quantum computing, which is very different from classical computers. However, many R&D teams have launched quantum computing software development tools (such as IBM's Qiskit, Google's Cirq, etc.) to make development more convenient; On the other hand, whether it is an algorithm or an application, because to convert into instructions that quantum can run, quantum computers need special editors to compile software;

6. Different application fields: The industry generally believes that the current quantum computer, compared with classical computers, will be more widely used in the following fields: material and drug research and development, financial portfolio optimization, logistics scheduling, machine learning training, weather prediction, etc., and classical computers have been widely used in all fields, driving the digitalization of the entire society.

7. Different levels of technological maturity: since the first electronic computer came out in 1946, after more than 60 years, the technology, process and industrial chain have matured, which can be made as small as a smart watch, and tens of billions of transistors can be integrated on a chip; The engineering research and development of quantum computers began from around 2012, and it has only been 10 years now, and the quantum computers announced so far are more like prototypes and technology verification machines, and the technologies, processes, and industrial chains are still unstable, let alone perfect.

The larger the "quantum volume",

Are quantum computers more powerful?

First of all, answer the question directly: quantum volume is indeed an important indicator to evaluate the comprehensive performance of subcomputers.

Why is there the concept of "quantum volume"? This is a hardware-independent indicator introduced by the industry in order to more accurately describe the performance of quantum computers with different technical routes. The quantum volume represents the maximum size of a square quantum circuit that a quantum computer can successfully achieve. In layman's terms, this performance index is related to the number of qubits, the maximum depth of the quantum circuit, fidelity, connectivity, crosstalk and other physical indicators. In general, the larger the quantum volume of a quantum computer, the more complex the problem it can solve.

"The Wandering Earth 2", released at the beginning of the year, made quantum computers popular again, and the concept of quantum volume also appeared in the film. But it is worth mentioning that although MOSS is very sci-fi, its nominal quantum volume of 8192 is obviously too small. In October 2022, quantum computing company Quantinuum has achieved the goal of 8192 in a 20-bit ion trap quantum computer, but quantum computing prototypes are still a long way from practical use.

As a performance indicator, quantum volume shows a certain degree of science and versatility at this stage. However, with the rapid development of quantum chip technology, there may be more relevant evaluation indicators in the future. For example, IonQ has introduced an indicator of "algorithmic qubits" to replace quantum volume. In the process of the gradual realization of quantum computers, it is foreseeable that more evaluation indicators that are more adapted to the evolution of technology will appear.

quantum computer,

Is it really "omnipotent"?

The word "omnipotent" is somewhat sci-fi, more suitable to describe the versatility and ubiquity of intelligent computers like the intelligent computer in "The Wandering Earth 2", which can be widely used in all areas of our work and life, as far as completing complex scientific calculations, warning of the Jupiter crisis, medium to analog digital life maps, close to chatting with us, helping us turn on the lamp in the room...

But quantum computers are not a panacea. Although from a realistic point of view, the trend of future computers becoming more and more powerful is relatively clear, and the future computer services available at any time, anywhere, and anywhere can also be expected, but computers are still not "omnipotent". And this is determined by the theoretical basis of computers - mathematics. There is a branch of mathematics called computability theory and computational complexity theory, which tells us that many problems in the world cannot be solved by calculation, no matter what computing tools are used, quantum computers are no exception.

So since a quantum computer is not omnipotent, what can it do, or what kind of computing tasks can it be good at?

From a mathematical point of view, there is a class of problems in computational complexity theory called "BQP problems", which can theoretically be solved by quantum computer polynomial time. Typical BQP problems are: integer factorization problems, discrete logarithm problems, some problems in simulated quantum systems, and so on. From this point of view, in the field of problems that mathematics can recognize, the problems that quantum computers can solve are only part of it.

To summarize, quantum computers are more suitable for calculating a large part of the problems that are very computationally intensive if they are calculated by classical computers, even if the best algorithms are used. An example of this image is the "needle in a haystack" problem: the algorithm for solving this problem is very simple, that is, to find every corner of the sea where every needle can fall. But this problem is a huge amount of work, and it is necessary to search the entire vast sea, from the surface to the bottom of the sea. Similar problems include looking up a number in a phone book without indexing, finding your avatar in a large number of faces, and so on. Even so, problems such as BQP can already be mapped to many application scenarios such as financial portfolio optimization, logistics and transportation optimization. If quantum computers do solve these problems in the future, the contribution to society will be enormous.

It is worth mentioning that the research and development of artificial intelligence large model technologies such as ChatGPT, which is currently hotly discussed, may also be related to quantum computing. Although there is no definitive answer, quantum computing is expected to play a positive role in the development and improvement of large models in terms of model compression, training speed and model performance, which deserves further attention.

From a physical point of view, quantum computers use the special mechanical principles of microscopic particles to achieve calculations, so those scientific research and practical application problems that are also based on quantum mechanics are theoretically more suitable for quantum computers to solve. We can think directly, such as computational chemistry, molecular simulation, new material research and development, medical drug research and development, and so on. These problems are all in the study and use of material properties at the atomic and molecular level to develop atomic or molecular structures with special properties that do not exist in nature. Quantum computers are likely to be the better option for the future.

Let's briefly talk about what quantum computers are not good at. Due to the implementation principle of quantum computers, there is a major problem in the current research and development process of quantum computers: noise problems. Simply put, when operating quanta for calculation, the stability of the quantum itself, the interaction between quanta, and the manipulation action itself will produce unwanted interference from the computing task itself. These disturbances can directly affect the accuracy of the calculation results, and even cause the calculation to be interrupted and failed. Although scientists are constantly improving quantum decoherence and error correction techniques to improve the probability of accurate calculations, they still do not reach 100%.

Current quantum computers cannot directly calculate a result, but can only obtain a definite result through observation under a certain probability. It's like letting the current quantum computer calculate "1+1", only a certain probability can get a result of 2, let alone let the quantum computer calculate more complex problems.

While classical computers can make mistakes, classic chips already have the ability to automatically find errors and fixes, so you can end up with accurate results. Therefore, this kind of accurate numerical calculation problems and logical determination problems, the current level of quantum computers compared with classical computers, is not good at. However, the field of quantum computing is also developing error correction technology suitable for quantum characteristics, quantum error correction code and quantum fault tolerance theory can produce more perfect qubits with multiple imperfect qubits through ingenious mathematical design, and achieve more steps and achieve more accurate calculations through specific fault-tolerant quantum operations. Finally, as the error correction code adopted becomes more and more powerful, the error rate of the entire calculation process can finally be controlled to close to zero. In this way, the quantum computer can do exactly what we want to achieve.

How to create "omniscient and all-powerful"

"The computer of the future"?

In the last question, we discussed that computers are not omnipotent, but with the current scientific and technological strength of human beings, it is not completely impossible to create a powerful "future computer" with the ability to surpass human knowledge and wisdom, solve complex scientific problems, and even support a huge virtual world system such as the "meta-universe". Just as human beings invented the first steam engine hundreds of years ago, and then gradually iterated technology and organizational forms, and established industrial production lines and industrial chains, the future computer manufacturing must also closely follow the evolving computing science and technology, and gradually practice, iterate and research and development.

"Good people do what they are good at", this is a relatively clear consensus in the industry for the future development of computers, in other words, to assign different computing tasks to the computing units and computing systems that are good at completion. As mentioned earlier, classical computers and quantum computers have their own computing tasks, so if you imagine the manufacturing ideas of future computers, a very important measure is to create quantum computers, and combine quantum computers with classical computers to integrate systems and even connect into computing power networks.

Making quantum computers is similar to classical computers, but also hardware systems and software systems must be created first.

The hardware system for manufacturing quantum computers is relatively difficult because it is still in the early stage of research and development. First of all, it is necessary to manufacture the core computing unit of quantum computers - quantum processors (QPU, also known as quantum chips). QPU is a hardware system that prepares and holds the smallest computing unit "qubit" for quantum computing, and functions like a CPU. At present, there are at least 7 technical routes for the preparation of qubits, which are adopted by the head quantum computer research and development institutions, including superconducting quantum, optical quantum, and ion trap, as well as neutral cold atoms, quantum dots, diamond NV color centers, topological quantums, etc., which are also under continuous development.

Secondly, with QPU, it is also necessary for QPU to run, generate quanta and manipulate quantum for computation. So a series of supporting systems are also required. Common supporting systems are: measurement and control systems, cryogenic equipment and components, ultra-high vacuum chambers, lasers, photon detectors and so on. QPU and supporting system form the "host" of quantum computers.

Third, the input and output of quantum computers currently need the assistance of classical computers, so the host of quantum computers is usually connected to several classical computers, which is convenient for operators to enter computing tasks through keyboards and mice, or import computing tasks from other computer systems.

Quantum computers also need software. First, a series of underlying software, similar to the current "motherboard driver" and "operating system", is needed to manage and control quantum computer hardware, compile and run quantum computing software; Secondly, "programming languages and tools" are needed to write quantum software and algorithms, because the principle of quantum computers is different from ordinary computers, so special languages and tools are also needed; Third, in order to solve practical problems in reality, quantum computers need to install corresponding quantum software, such as quantum chemical analysis software, quantum simulation software, etc.; Fourth, inside the quantum software, the relevant quantum algorithms are generally loaded to complete the calculation work;

QPU, supporting systems, input and output systems, underlying software, programming languages and tools, applications, algorithms, these hardware and software systems need to operate together at the same time in order to truly exert the excellent computing power of quantum computers and create value for society.

The scientific research and exploration of combining quantum computers with QPU as the computing core and classical computers with CPU+GPU as the computing core has begun in the field of high-performance computing. High-performance computing is the pursuit of the ultimate computing power in the field of computers, which leads the development of the entire computer industry, and many computer industry technologies, such as Hadoop, are first developed by the field of high-performance computing. Internationally, such as IBM's quantum computing team, quantum computers have taken quantum computers as the core of their next generation of high-performance computing, and the "European Joint High Performance Computing Project (EUroHPC JU)" has launched experimental research on the integration of quantum computers and supercomputers; Chip company NVIDIA also recently launched QUDA, a hybrid computing platform that can integrate quantum computing and high-performance computing resource calls; Cloud computing service providers such as Amazon and Microsoft have also seen this trend, put the capabilities of quantum computers to the cloud, and have successively opened up quantum computing cloud platform services. The era of high-performance computing 3.0 of CPU+GPU + QPU has begun.

It is believed that with the continuous progress of quantum computer research and development, as well as the combination and collaborative development of quantum computers and classical computers, a powerful "future computer" will definitely come out in the near future, and assume decisive tasks and responsibilities at similar human historical moments in "The Wandering Earth 2".

Does quantum computing have practical applications now?

In recent years, along various technical routes, the research and development of quantum computer prototypes have made great progress. Although quantum computers at this stage are still far from being truly practical, the current level of engineering has far exceeded the most optimistic predictions of scientists 20 years ago. Looking back on the past 2022, many new development achievements have emerged in the field of quantum computing prototypes, which can be described as a hundred flowers blooming and a hundred birds contending.

Among them, the superconducting quantum computer route is the most dazzling, as always leading. At present, most of the domestic and foreign technology companies have chosen the superconducting route for long-term layout and investment, and they all have good report cards in 2022. IBM released its 433-qubit quantum computing chip as scheduled, which is currently the highest level in the global superconductivity direction, and is expected to reach 1,000 qubits in 2023. Google pays more attention to the quality of qubits and has made continuous progress in quantum error correction.

The ion trap route is characterized by a particularly long coherence time of bits, so it is possible to prepare a very high-fidelity quantum logic gate (somewhat similar to a logic gate composed of transistors on a chip), so the gate fidelity is very high. Although ion trap quantum computers face more difficulties in increasing the number of bits (~20 bits), their unique advantages can be seen by the previously discussed indicator quantum volume (QV). In February 2023, leading company Quantinuum announced that based on its H1 series of quantum chips, the quantum volume can reach 32768 (2 to the power of 15), setting a new record.

The prototype of the optical quantum computing route has always been an important carrier to demonstrate the superiority of quantum computing. In June 2022, Xanadu completed a Gaussian boson sampling experiment based on its programmable optical quantum prototype, demonstrating the superiority of quantum computing once again. The neutral atomic route is the dark horse in the technological development map in 2022. In September 2022, Pasqal France announced the launch of a 324-atom (qubit) quantum processor. But its usability at the engineering level has yet to be verified. The semiconductor quantum computing route has also faced difficulties in 2022 and made a lot of progress. Intel updated the number of silicon spin qubits to 12 in October 2022, while the yield of chip production is as high as 95%, which is a key step towards commercialization.

Now that quantum computing prototypes have made so much progress, do quantum computers now have some practical applications that can be implemented? In fact, there is still a long way to go before quantum computers can actually produce commercial applications. Overall, the current quantum computing prototype is in the early stage of quantum computing development, and is still in the era of NISQ (Noisy Intermediate-Scale Quantum) in the field of quantum computing. In order to achieve the operation of iconic quantum algorithms and applications such as Shor's algorithm, a large number of researchers still need to work hard to achieve quantum error correction and achieve million-scale quantum bit coherent manipulation. This process is generally predicted to take at least 10-15 years of scientific research innovation and engineering technology accumulation.

So are quantum computers in the NISQ era useless? Not really. Through quantum error mitigation, classical-quantum hybrid and other schemes, scientists are also very active in exploring the application potential of quantum computing prototypes at this stage, which are used to explore the possibility of solving problems in fields including finance, chemical industry, biology, medicine, aviation and artificial intelligence.

Typical application scenarios in the financial sector include portfolio optimization, high-frequency (quant) trading, and fraud detection. Typical application scenarios in the chemical industry include the chemical industry and catalyst design. In aviation, the goal is to use quantum algorithms to help urban air mobility develop more flight routes and improve the rescheduling of rail traffic lines. In the field of artificial intelligence, the goal is to use quantum algorithms to speed up the learning time of robots and use the latest machine learning technology.

quantum computing,

Is it better to do simulation?

Quantum computing can simulate the evolution of quantum mechanical systems faster than classical computers, which is indeed the original motivation of Feynman, the precursor of quantum information science, to propose the idea of quantum computing. The use of quantum computers to simulate quantum systems is also currently considered one of the most promising applications for quantum computing, and may be one of the first areas of industrial relevance.

Let's take a look at three examples: materials science, drug discovery, and protein folding.

Quantum computing can effectively simulate quantum evolution and has important applications in materials science. For example, in material chemistry simulations, once the spin number is greater than about 100, some magnetic models cannot be simulated on classical computers at all. However, it is estimated that such a system can be simulated in about an hour using a quantum computer with 500,000 qubits. For more complex systems, more qubit resources may be required. However, compared with traditional computing methods, the introduction of quantum computers has greatly expanded the boundaries of scientific problems that can be explored, and it is expected to be applied to the development of novel characteristic materials in the future.

The application of drug discovery is also an active area. In principle, quantum computers can accurately simulate the properties of small drug molecules that simply cannot be simulated using traditional computational methods. At this point, it can be said that quantum computers have the advantage. However, whether this advantage is large enough to be useful to drug development companies in practice is still under active research.

For example, if a molecule can't be modeled at all using classical computers, but can be modeled in a matter of months using millions of qubits, does that have practical advantages? If this drug can cure serious diseases, it is indeed an advantage in a sense. Quantum hardware and algorithms also have the potential to improve to the point where the time consumed by the computational process and the qubit requirements are significantly reduced.

Quantum computing is also being studied to model the way larger molecules such as proteins fold, a notoriously difficult problem to solve on classical computers, even using breakthrough techniques in machine learning such as AlphFold. Again, this is known to be an area where quantum computers have a theoretical advantage, although the problem of practical superiority is similar to that of small molecules.

What are there at present

Companies and institutions doing quantum computing?

According to the statistics of McKinsey's "Quantum Science and Technology Observation", the total number of enterprises and institutions engaged in quantum computing at home and abroad has exceeded 450, respectively, in quantum chips, quantum computer systems, quantum algorithms, quantum software development tools, quantum application software, quantum cloud platform and other aspects of research and development. Enterprises and institutions engaged in quantum computing can be roughly divided into four categories: head information and technology enterprises, national scientific research institutions, start-ups and scientific research institutions.

Leading information and technology companies, such as IBM, Google, Microsoft, Intel, Honeywell, NVIDIA, Amazon, Hitachi, NEC, Alibaba, Huawei, Baidu, Tencent, etc.;

National scientific research institutions, such as Fermi National Laboratory and Argonne National Laboratory in the United States, and the Institute of Quantum Information and Quantum Science and Technology Innovation of the Chinese Academy of Sciences in China, etc.;

Most of the start-ups were established after 2015, and the more representative ones are D-Wave, Rigetti, Xanadu, IonQ, Quantinuun, Origin Quantum, Guodun Quantum, etc.;

Scientific research institutions, such as MIT, MIT, Massachusetts Institute of Technology, Harvard University, Marx Planck Society, University of Science and Technology of China, University of Tokyo, University of Cambridge, etc.

Before quantum computing is actually applied, there is

What needs to be done? How can it

Accelerate real-world adoption?

Quantum computers still have a long way to go before they can be actually applied.

From the perspective of the system, in terms of hardware, it is necessary to improve the quantity and quality of qubits, improve the level of quantum error correction technology and manipulation observation technology, and complete the standardization of quantum computer manufacturing process, optimization of operating costs, and optimization of operating energy efficiency. In terms of software, it is necessary to develop quantum algorithms with practical application value and develop quantum application software; In terms of theory, it is necessary to complete rigorous mathematical proofs of computational superiority, a large number of experiments and scientific demonstrations of economic value, and so on.

From a generational point of view, quantum computers will also go through the NSQ era of "noisy medium-quantum scale quantum computers", and eventually reach the FTQC era of "fault-tolerant quantum computers" with sufficient quantum quantity and quality.

Source: Information and Communication Technology and Policy Issue 7, 2020, Quantum Computing Technology Industry Development Status and Application Analysis

From the overall perspective of quantum computer research and development, there seems to be no shortcut to accelerate the practical application of quantum computers, except for more determined research and development, practice and reflection, and forging ahead. The ancient adage said: "If you don't accumulate small rivers, you can't make rivers and seas." If you don't accumulate steps, you can't go a thousand miles. "The research and development and application of science and technology cannot be rushed.

Especially for cutting-edge exploratory technologies such as quantum computing, long-term cooperative research and experimentation between the scientific and technological community and the industrial community are more needed, and even repeated trial and error. The maturity and popularization of classical computer technology has also gone through 40 years; The development of quantum computers has only begun for 20 years, which is not long.

Take IBM as an example, IBM has been conducting quantum computing research and development for 12 years and is one of the earliest pioneers to invest in the development of quantum computers. In 2011, IBM began to conduct theoretical and experimental research on quantum computers; In 2016, IBM put quantum processors on the cloud for the first time to provide experimental services; In 2020, after investing in quantum computing research and development for 10 years, IBM announced its own research and development roadmap from 2019 to 2023, moving forward more determinedly; In 2022, IBM updated its roadmap to 2025 and adjusted its technology roadmap beyond 2022.

It can be seen that IBM is likely to achieve technology updates after a large number of experiments, trial and error, and even commercial considerations; It is gratifying to note that since 2020, IBM has been firmly fulfilling its announced roadmap and milestone plan, and continues to make key contributions to the development of quantum computer technology. In December 2022, IBM delivered the Osprey superconducting quantum chip that can prepare 433 qubits as scheduled, marking a solid step forward in the research and development of superconducting quantum computers.

Another example is Google, which also began to invest in quantum computer research before 2015. In the past 8 years now, the industry and even the whole society, the understanding of quantum computers is constantly increasing, and even overhype has appeared from time to time, and many people even think that quantum computers will be commercially available within a few years. It was not until this year, that is, February 2023, that Google published its progress in the field of quantum error correction in the journal Nature, initially proving through rigorous scientific experiments that quantum error correction can indeed improve the accuracy of quantum computing, and this is only the second step for Google to implement a usable quantum computer. There is still a long way to go before Google's plan to achieve a truly commercial quantum computer in 2029.

From the perspective of the NISQ era in which the current quantum computer is located, under the conditions of existing quantum computer hardware, finding more algorithms with both computing superiority and practical economic value, as well as quantum application software for practical application scenarios, is considered to be an important idea to accelerate the practical application of quantum computing.

Taking quantum algorithms as an example, in order to do this, on the one hand, academia and industry need to cooperate together, constantly practice how to reasonably explore and deploy quantum algorithms in real scenarios such as materials, pharmaceuticals, finance, physics, etc., and eventually reach the level of ability to surpass classical computing in some scenarios. On the other hand, at this stage, more deployment of quantum computer hardware and software platforms open to the public can accelerate the industry to develop, deploy and test quantum algorithms, and quickly screen out potentially valuable applications. This process will also feed back the development of quantum hardware and software applications.

Quantum computing is an epoch-making hard core technology that is expected to become an important cutting-edge science and technology that builds future computing power and reshapes the future world. With the continuous investment in quantum computer research and development at home and abroad, we will cross the NISQ era and enter the FTQC era, and further usher in the emergence of a large number of quantum algorithms and applications with both computing superiority and practical economic value. We are looking forward to the future.

For the present, we still need rational optimism, objectively understand quantum computers, and fully understand its process from technology research and development to practical application. More importantly, the development of the field of quantum computing requires more active cooperation among enterprises, universities, research institutions and other parties in the field of quantum computing at home and abroad, do a good job in research and development in a down-to-earth manner, and solve various scientific and engineering challenges step by step.