Abstract Quantum computing and quantum computers have attracted much attention from both the academic community and industry in recent years. By exploiting the quantum properties of materials, scientists are aiming to overcome Moore's law of miniaturization and develop novel quantum computers. The concept of quantum computing was first introduced by the distinguished physicist Richard Feynman in 1981. As one of the early pioneers in this field, Turing Award laureate Andrew Chi-Chih Yao made a seminal contribution in developing the theoretical basis for quantum computation in 1993. Since 2011, he has served as the founding director of Tsinghua University's Center for Quantum Information (CQI), which aims to become a world-class research center for quantum computing. In a recent interview with NSR, Yao recounted the history of quantum computing and expressed his view on the future of this field. He suggests that quantum computers could excel in many tasks such as the design of new materials and drugs as well as in the simulation of chemical reactions, but they may not supersede traditional computers in tasks for which traditional computers are already proven to be highly efficient. NSR: Quantum communication and quantum computing have attracted much attention in the media. Aren’t these two different concepts? Yao: These are indeed two different concepts, although they are related. The techniques used in quantum computing are more advanced. The main motivation of quantum communication is the development of cryptography for confidential communication. The signals to be transmitted from one place to another do not need to be highly precise. But quantum computing requires high precision of the signals. In the past decade or so, a few emerging technologies have been developed by major companies, such as Google. It is the general feeling that usable technology will appear in five to six years. The theoretical foundation was already established 20 years ago; now it is the problem of implementation. CHALLENGE FOR MOORE’S LAW OF MINIATURIZATION NSR: Quantum computing has become a hot topic. What is its basic principle? Yao: Rapid progress in the miniaturization of semiconductor circuits has made computers increasingly powerful. However, miniaturization has an intrinsic limit—when the size of the circuit elements on a chip is reduced beyond the nanometer scale, quantum mechanical effects become dominant and may hamper the elements’ proper functions. This is the end of Moore's law of miniaturization. This is an unavoidable destiny for traditional computers, but scientists began to think whether one can convert the quantum phenomenon from an enemy into a friend—to construct a computer that computes by the logic of quantum mechanics, as described by Schrodinger equations, rather than Boolean logic. The idea of a quantum computer was first proposed by Richard Feynman in 1981. He said that one could in principle design a computer that operates by the characteristics of quantum mechanics, simulates quantum systems and obtains solutions using quantum equations. Feynman's idea has attracted much attention in the academic field. View largeDownload slide Chi-Chih Yao, Dean of Institute for Interdisciplinary Information Sciences, Tsinghua University (Courtesy of Chi-Chih Yao). View largeDownload slide Chi-Chih Yao, Dean of Institute for Interdisciplinary Information Sciences, Tsinghua University (Courtesy of Chi-Chih Yao). Traditional computers perform their functions using two-value Boolean logic (0 and 1), through integrated electronic circuits. The computation maps the input points represented by the bits to a higher level and, after multiple mapping, reaches the output points to provide the final solution. The qubits of a quantum computer, however, can represent one, zero, or any superposition of these two states. The computation of a quantum computer system is analogous to the rotation of solid objects; the result of the computation is read off by measurements. Compared to traditional computers, where one operation involves one defined path, the operation of quantum computers can proceed along multiple computational paths to reach the goal, because quantum wave functions dictate the existence of multiple states at the same time. This phenomenon is known as quantum parallelism. Quantum parallel computation is the key reason that quantum computers can attain much faster computing speeds than traditional computers. NSR: What are the main differences in hardware design between traditional and quantum computers? Yao: Quantum computers are relatively closed systems; computation can be done almost instantaneously. Basically, they are very ‘shy’: once observed, the computation is interrupted and stopped. Furthermore, quantum computers are very sophisticated systems involving multiple cutting-edge technologies. For example, the storage unit of the computer, communication between multiple units and modulation of qubit states all require the use of lasers. The materials and manufacturing processes required for making quantum computers represent the integration of many advanced technologies developed over the past three to four decades, and involve close interdisciplinary collaboration. NSR: Does the uncertainty associated with the quantum phenomenon affect the accuracy of quantum computation? Yao: Yes, but an uncertain answer is not necessarily a wrong answer. Some quantum computation could in fact always yield correct answers. But for practical computation, some error is acceptable, without requiring 100% accuracy. FROM THEORY TO PRACTICE NSR: Although the concept of the quantum computer appeared in the early 1980s, it seems that progress was slow in the subsequent decade. Yao: Indeed. After Feynman raised the idea, it was mainly physicists who made further theoretical exploration. By the early 1990s, physicists had mostly figured out how quantum computers should operate and computer scientists began to step into this field—I myself was among the earliest to do so. In 1994, Peter Shor of Bell Lab devised a quantum computing algorithm for breaking the cryptogram, triggering a wide interest from the computing community, and the US government and NASA began to invest heavily in this field. Research teams began to appear, in competition for making the first realistic quantum computer. NSR: What are the major advances since then? Yao: Major advances were made in the exploration and choice of approaches for realizing quantum computers. Over the last decade or so, scientists have tried various materials, such as ion traps, superconductors and diamond for constructing the quantum computer. Recently, topological insulators also entered the scene, because of their excellent correctability function. But a long way still lies ahead. One of the main difficulties is to keep the ultra-low temperature of the functional state. NSR: When do you think the first quantum computer will appear? Yao: Many people predict that the first quantum computer will appear in five or six years, but I think to make a quantum computer that is capable of reliable computation over thousands of qubits is not easy. Large companies such as Google and IBM have also invested heavily in research into quantum computers. In particular, Google has recruited one of the most important experts in this field, John Martinis, and his entire team from UC Santa Barbara, and they are making good progress. NSR: Under your initiative, Tsinghua University established the Center for Quantum Information (CQI) in 2011. What is your goal? Yao: Our goal is to build a world-class center for research in quantum information and for training the next generation of scientists in this area. Thus, our immediate step was to recruit high-quality researchers, such as Professor Luming Duan, a physics professor at University of Michigan, through the 1000 Talents Program. He has done excellent work in our center in the past few years. The work by his team on diamond color-center quantum computing is now pioneering in the field. They were the first to realize universal geometric quantum gates with solid spins [Zu C, Wang W-B, He L et al. Nature 2014; 514: 72–5] and made the world's current largest diamond-based quantum computation system (11 bits). NSR: What is the advantage in using the diamond system? Yao: There are two advantages: first, it can work at room temperature; second, it has a solid-state crystal structure, which could be extended to larger scales if it works well for a few qubits. In addition to the diamond system, our center is also doing research and making good progress in the use of ion traps, superconductors and photon networks. BEYOND QUANTUM COMPUTATION NSR: It is exciting to know the super-high capability of quantum computers. Will they replace traditional computers in the future? Yao: I think that traditional and quantum computers will coexist. Each has their own advantages. Traditional computers have the precision and mature technology that quantum computers do not have yet. But quantum computers will have an I think that traditional and quantum computers will coexist…. Quantum computers are best applied to situations where quantum mechanical effects dominate, excellent for solving problems in the fields of material design, drug development and physical chemistry. —Chi-Chih Yao advantage over classical computers in tackling problems involving the quantum mechanical effect. For example, they would be excellent for solving problems in the fields of material design, drug development and physical chemistry, which are most difficult to solve by traditional computers. NSR: Both the hardware and software of quantum computers are quite different from traditional computers. What are the main challenges at present? Yao: Quantum computing is a typical interdisciplinary area requiring close collaboration among scientists and engineers, especially between quantum physicists and computer scientists. A breakthrough in algorithm development will stimulate improvement of hardware, and vice versa. For example, Peter Shor, whom I mentioned earlier, not only proved that quantum computing can solve the problem of decoding cryptograms, but also solved the problem of error correction in quantum computing. It was his research that convinced physicists of the feasibility of the quantum computer. When the quantum computer develops to a certain stage, a major transformation of computer science will be required. Data storage, operating systems and the programming language of traditional computers will need to be redesigned. It is not clear how this could be done at present, but it is a significant research direction. Many leading IT companies have already established substantial programs to develop quantum software. Research on quantum computing methods and algorithms is an area of great potential. Several elegant computing methods have appeared in the past decades, all fascinating theoretically. I hope to see the development of more quantum computing methods that are linked to practical use, such as methods for designing materials. NSR: Quantum computers appear to require development in both frontier science and high-level manufacturing technology. Yao: Indeed. I have always stressed that the significance of developing a quantum computer in China is far beyond just doing quantum computing—it will drive the development of related technologies. Similar to the lunar exploration program in China, such a project will activate the potential of scientists and engineers and, through solving specific problems, invent new methods and technologies that will benefit society at large, including industrial development and national security. Seventy years ago, China missed the opportunity to occupy the high ground in microelectronics; hopefully we will not miss out this time in developing quantum computers. Many countries in the world are now investing heavily in quantum computing, in order to compete for the future leading position in this field. We have immediate concerns when doing actual experiments. For example, diamond material suitable for quantum computers depends on imports from abroad. Other countries could easily refuse to sell the material to us, especially when the situation becomes increasingly competitive. If we do not settle down to begin developing these materials ourselves, we will be easily constrained in the future. Work in this area may not immediately lead to publications useful for personal evaluation and promotion. Unless our evaluation system is changed, it will be difficult to motivate researchers to do this groundwork. Thus, we remain dependent on the importation of materials and technologies at present. THE YAO CLASS AND INNOVATIVE EDUCATION IN COMPUTER SCIENCE NSR: You have made enormous efforts in improving computer science education in China. The ‘Yao Class’ has become a model for educating talented young students. How did the ‘Yao Class’ come about? Our new teaching approach [of the ‘Yao Class’] emphasizes discovering the students’ talent and potential. We put together a curriculum with up-to-date and challenging courses, taught by the best professors that we could find. They also motivate the students with research problems to stimulate their interest and imagination. —Chi-Chih Yao Yao: In 2005, I had an in-depth conversation with Harry Shum, then-director of the Asia Institute of Microsoft, on the higher education of computer science in China. We both felt that there are many excellent students in China, but most of them fall behind their counterparts in the US because of poor curricula and education methods. Even the best students who pursue higher degrees in the US always have to go through a painstaking transition period, due to poor training for independent and in-depth thinking in their undergraduate years. Another reflection of this deficiency is that in the US the number of tenured professors in computer science from China is much smaller than that from India. After extensive discussion and careful consideration, we decided to develop a new model to nurture promising undergraduate students in the field of computer science at Tsinghua University. NSR: How do you select students for the Yao Class? View largeDownload slide Professor Chi-Chih Yao giving a lesson to the Yao Class (Courtesy of Chi-Chih Yao). View largeDownload slide Professor Chi-Chih Yao giving a lesson to the Yao Class (Courtesy of Chi-Chih Yao). Yao: We hold an optional entrance exam for the freshman class in Tsinghua University each year, and about 30 students are selected; most of them are top students in the freshman class or medalists in the Mathematics, Physics and Information Science Olympiads. NSR: What's special about the Yao Class curriculum? Yao: First, we strengthen the courses on computer algorithms. These are required for undergraduate computer science majors in the US, but mostly absent in Chinese universities. Second, we optimize the curriculum by focusing on the quality rather than quantity of the courses. Undergraduates in China tend to take lots of courses to get many credits, without studying any subjects in depth. Our new teaching approach emphasizes discovering the students’ talent and potential. We put together a curriculum with up-to-date and challenging courses, taught by the best professors that we could find. They also motivate the students with research problems to stimulate their interest and imagination. Finally, all courses in the Yao Class are taught in English, and the Class provides opportunities for all students to study abroad for a semester. Thus, many students are capable of asking good questions in their first two years in the class, and capable of further studying on their own. By the third and fourth years, they have acquired the ability to do research, even reaching the level of graduate students. This is something we are rather proud of. NSR: That is very impressive; qualified instructors are also very important, aren’t they? Yao: This in fact may be different from other universities. Most professors of the Yao Class are young; they are creative and enjoy interacting with the students. Beyond teaching, they also guide our undergraduates in venturing into research. NSR: What about the teaching of graduate students? Yao: In graduate school, the academic training is focused on research. Professors mostly offer advanced courses related to their own research fields. Students are expected to learn and work with their thesis advisors. In this regard, the way of training graduate students is pretty much the same all over the world. The quality of the graduate program faithfully reflects the quality of the institute. We have conscientiously built up a high-quality faculty in our Institute (Institute for Interdisciplinary Information Sciences, IIIS) and cultivated a strong reputation in education and research. INTERDISCIPLINARY FUTURE OF COMPUTER SCIENCE NSR: Interdisciplinary research and boundary-crossing innovation is an emerging trend in the world. Is this true for computer science? Yao: Yes, this is inevitable. Foreseeing the trend, we established the IIIS at Tsinghua University in 2011, perhaps one of the first of its kind in the world. You see, computer science used to be a somewhat introverted discipline in the past three to four decades, focusing mainly on internal problems within the area of the computer itself, and had accumulated a large amount of theories and knowledge. The situation has changed in the past decade. With the rapid development of the Internet, many techniques developed within computer science suddenly became very useful. For example, there is a branch in computer science called distributed system computing, which can ensure the robustness and accuracy of computing systems. A few years ago the idea of the blockchain became popular and widely used in the financial system. In essence, it is an improvement and application of distributed system computing. This is a golden time for computer scientists—their research can have a great impact on the world beyond the ivory tower. This will also change our criteria for the evaluation of research accomplishment. NSR: A similar change is happening in many other fields as well. Neuroscience, for example, has become more intertwined with computer science. The China Brain Project, one of the major frontier science projects for 2030, will soon be initiated as a project on ‘brain science and brain-inspired intelligence technology’. It will have a large component in the research on the development of new computing methods and algorithms based on the principles of information processing in the brain. Yao: That is very exciting and promising. Brain science involves multi-level and multi-faceted explorations. Both brain science and computer science are concerned with the problem of learning. Although AlphaGo has beaten the best human Go player, we are still far from understanding how the human brain learns. This difficult problem may be solvable by the fusion of research in these two areas. NSR: That is indeed one of the goals of the China Brain Project. Brain-inspired intelligence research stresses the integration of features of the brain into machine learning, aiming to transcend the supervised learning models of most current paradigms such as deep learning networks that require large data sets and computing power. Yao: It is also a challenge that scientists from many different disciplines want to address, and that is why we need interdisciplinary research and cross-disciplinary cooperation. PERSONAL RETROSPECTIVES NSR: We are also interested in your own experience: you have always remained at the forefront of computer science; how have you done that? Yao: Before I turned to computer science, I had been studying physics and obtained a PhD degree in physics from Harvard. Physics is a problem-driven science, helpful for developing the habit of open thinking. When I stepped into the field of computer science, I would first sort out the core issues and then try to discover appropriate approaches to resolve them. When I stepped into the field of computer science, I first sorted out the core issues and then tried to discover appropriate approaches to resolve them. —Chi-Chih Yao After I received another PhD degree in computer science in 1975, I began to work on computing theory and complexity problems, and made some accomplishments. Around 1980, when personal computers and networks began to change the computational approaches, I realized that network security and cryptography were important issues. I and others then established complexity-based cryptography, which is a shining achievement of computer science. Around 1990, I became aware that physicists were working on quantum computing. I found this very interesting. It led me to start thinking about quantum computing and quantum communication and I made some contributions in these new directions. NSR: What is your most recent research interest? Yao: Besides quantum computation, I have become interested in computational economics, in particular auction theory, which deals with how people act in auction markets and the properties of auction markets. In the past 10 years, computer scientists have joined economists to make auction theory a new interdisciplinary area. By designing and analyzing auction mechanisms and strategies, we can formulate the best market rules and auction mechanisms for achieving specific purposes. NSR: It is widely known that you recently acquired Chinese citizenship and gave up American citizenship. Why did you make this decision? Yao: It was a natural and spontaneous decision, I feel very happy to be a 100% Chinese citizen. I have been working at Tsinghua University for nearly 14 years, and relish the sense of achievement made together with my colleagues and students here in China. Mu-ming Poo is director of the Institute of Neuroscience, Chinese Academy of Sciences and Executive Editor-in-Chief of NSR, Ling Wang is a freelance science writer based in Beijing. © The Author(s) 2018. Published by Oxford University Press on behalf of China Science Publishing & Media Ltd. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)
National Science Review – Oxford University Press
Published: Apr 26, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera