Milestone breakthrough! “Nine chapters” of Chinese quantum computing prototype came out to realize “quantum hegemony”

On December 4, a research team composed of Pan Jianwei and Lu Chaoyang of the University of science and technology of China, cooperated with Shanghai Institute of Microsystems, Chinese Academy of Sciences and National Research Center for parallel computer engineering technology, built a prototype of 76 photons “nine chapters”, and realized the fast solution of the “Gaussian boson sampling” task with practical prospects. < p > < p > according to the existing theory, the quantum computing system can process Gaussian boson sampling by 100 trillion times faster than the fastest supercomputer at present, that is, the task of “nine chapters” can be completed in one minute, and the supercomputer will take 100 million years. This achievement makes China successfully reach the first milestone of quantum computing research: the superiority of quantum computing. The related papers were published online in international academic journals on December 4. Due to the ultra fast parallel computing ability of quantum computer in principle, quantum computer can realize exponential acceleration compared with classical computer in some problems of great social and economic value. At present, the development of quantum computer has become one of the biggest challenges of the world’s scientific and technological frontier, and is the focus of competition among developed countries in Europe and the United States. The first stage goal of quantum computing research is to realize the “superiority of quantum computing”, that is to develop a prototype of quantum computing, which surpasses the classic supercomputer in solving specific tasks. < / P > < p > the “specific task” of quantum computing mentioned above refers to the problem that has been carefully designed and is very suitable for quantum computing devices to realize their computing potential. Such problems include random quantum circuit sampling, IQP circuit sampling and Gaussian boson sampling. Google’s quantum AI team focuses on random quantum circuit sampling. < / P > < p > the Galton board problem was put forward by British biostatistician Gordon. This problem can be understood as the ball is dropped from the top. Half of the ball may walk from the left and half may walk from the right when passing through a nail board. When many small balls fall randomly from top to bottom, the distribution of the number of small balls falling in the lower grid will be certain This model can be used to intuitively understand the central limit theorem. < / P > < p > if we develop a quantum version of the “Galton nail board”, that is, the ball is replaced by identical photons, and nails are replaced by beam splitters, then the game will become a quantum simulation of “Bose sampling”. Generally speaking, “Bose sampling” refers to the sampling of the entire output state space of n identical bosons after passing through an interferometer. < p > < p > computer scientists s.aronson and a.arkhipov proposed a fast method to calculate the constant value of matrix in 2013. The main principle is to sample and analyze the probability distribution of bosons processed by linear devices, so as to quickly find a constant value method of N x n dimension matrix. Particles in nature are divided into bosons and fermions, and photons belong to bosons, so we can use photons to realize boson sampling experiments. From the point of view of computational complexity, the number of solving steps increases exponentially with the increase of photon number. For such a difficult problem, supercomputers can be defeated in small and medium scale. Therefore, the problem of Bose sampling has been focused on by scientists in the field of quantum computing. They are ready to challenge classical computers with it. < / P > < p > again, why is it difficult to simulate random quantum circuits in classical computers? For example, for a 50 bit random quantum circuit sampling, the dimension of the state space of the final output quantum state is 250. If the classical computer simulation is used, it is extremely difficult to first store such a high-dimensional quantum state. Secondly, in such a high-dimensional computing space, the quantum computing operations of each layer are simulated until the final calculation result is output It’s even more difficult! In recent years, it has been recognized that there are two ways to demonstrate the advantages of quantum computing, i.e., random circuit sampling using superconducting qubits and boson sampling by photons. On the second route, the team of CUHK has always been leading the world. In 2019, they have realized Bose sampling quantum computation with 20 photon input and 60 × 60 mode interference circuit. The dimension of output state space is as high as 370 trillion, and its complexity is equivalent to 48 quantum bits, which is close to the “superiority of quantum computing”. < / P > < p > this time, pan Jianwei’s team independently developed a quantum light source with high efficiency, high homogeneity, extremely high brightness and large-scale expansion ability. At the same time, it meets the requirements of 100 mode interference circuit with phase stability, fully connected random matrix, wave packet coincidence degree better than 99.5% and pass rate better than 98%, phase-locked accuracy within the negative 9th power of relative optical path 10, and high-efficiency 100 channel superconductivity Nanowire single photon detector has successfully constructed a 76 photon 100 mode Gaussian Bose sampling quantum computing prototype “nine chapters”, which means to commemorate the earliest mathematical monograph in ancient China. < / P > < p > “some of the key devices used in our experiments have been banned by foreign countries, but we have made the best quantum light source in the world by relying on ourselves and domestic cooperation units. After all, science is for all mankind. ” Pan Jianwei told reporters that to complete this experiment, more and better photons are needed. It is precisely because we have the best quantum light source in the world. The space scale of the output quantum state of this experiment has reached the 30th power of 10. < / P > < p > according to the current optimal classical algorithm, “nine chapters” can process Gaussian boson sampling by 100 trillion times faster than Fuyue, the world’s No. 1 supercomputer, and equivalent to 10 billion times faster than the 53 bit quantum computing prototype “Platanus” released by Google last year. At the same time, the superiority of quantum computing proved by Gaussian Bose sampling does not depend on the number of samples, which overcomes the loophole that quantum superiority depends on the number of samples in Google’s 53 bit random circuit sampling experiment. < / P > < p > this achievement firmly establishes China’s position as the first square array in international quantum computing research, and lays a technical foundation for the realization of a large-scale quantum simulator that can solve major practical problems in the future. In addition, the Gaussian Bose sampling algorithm based on the “nine chapters” quantum computing prototype has potential applications in graph theory, machine learning, quantum chemistry and other fields, and will be an important direction of future development. It is precisely because of the ultra fast parallel computing ability of quantum computer in principle that it is expected to achieve exponential acceleration in some problems of great social and economic value, such as cipher decoding, big data optimization, material design, drug analysis, etc. In fact, the development of quantum computers is a challenging and potentially long-term task. In order to promote the research and development of quantum computer, we must divide it into small goals and break through them one by one. The first small goal is “quantum superiority”, which means that the computing power of quantum computer on a specific problem far exceeds the supercomputing with the best performance, proving the superiority of quantum computer. Therefore, “quantum superiority” is regarded as an important milestone in the development of quantum computing. < / P > < p > the first stage is to develop a high-precision special quantum computer with 50-100 qubits. It can solve some specific problems with high complexity that supercomputers can’t solve efficiently, and realize the milestone of “quantum computing superiority” in computing science. < / P > < p > in the second stage, through the accurate preparation, manipulation and detection of large-scale multi-body quantum systems, we develop a quantum simulator that can manipulate hundreds of quantum bits coherently, which can be used to solve some practical problems that supercomputers are not competent for, such as quantum chemistry, new material design, optimization algorithm, etc. The third stage is to improve the manipulation accuracy of qubits by accumulating various technologies developed in the development of special quantum computing and simulator, so as to reach the critical fault-tolerant threshold of quantum computing. Many foreign professors in related fields, including many Wolfe prize winners, academicians of the American Academy of Sciences and other senior experts, have commented in succession, praising the great achievements made by Chinese scientists. Ignacio cirac, director of the Max Planck Institute in Germany and winner of the Wolff prize and Franklin medal, commented: “in general, this is a major breakthrough in the field of quantum science and technology, and a big step towards the development of quantum devices with quantum advantages over classical computers. Professor Pan’s team is unique in the world, and they have produced many significant results including this experiment. ” Anton zeilinger, President of the Austrian Academy of Sciences, academician of the American Academy of Sciences and Wolff prize winner, believes that this work is important because pan Jianwei and his colleagues have proved that photon based quantum computers can also realize “quantum computing superiority”. Dirk Englund, a professor at MIT and winner of the president’s Award for young scientists and Sloan award, said it was an epoch-making achievement and a remarkable achievement. This is a milestone in the development of these medium-sized quantum computers. Philip Walther, a professor at the University of Vienna and a member of the American Physical Society, commented: “in their experiments, they have obtained the calculation results that can only be given by the strongest classical computer in a trillion years, which provides a strong proof of the super power of quantum computers.” Barry Sanders, Professor of the University of Calgary and director of the Institute of quantum science and technology in Canada, commented: “I think this is an outstanding work that has changed the current pattern. We have been trying to prove that quantum information processing can overcome classical information processing. This experiment is beyond the reach of classical computers. ” Tim Ralph, a professor at the University of Queensland, commented: “I believe the paper by Professor Pan and Professor Lu is a major breakthrough. It’s a real “hero” experiment, pushing the technology of all aspects of the experiment far beyond the previous level. ” Peter zoller, academician of the American Academy of Sciences, winner of the Wolf Prize and Dirac medal, said that using quantum devices to solve increasingly complex problems and embody quantum advantages is one of the most important issues in the frontier of quantum science. The Gaussian boson sampling experiment based on photons carried out by Lu Chaoyang, pan Jianwei and colleagues has raised the research level to a new level in terms of the size and expansibility of quantum systems and the prospect of practical application. Yueshang group has become the third social e-commerce service platform listed in China after being gathered in pinduoduo