Quantum computer is a device that uses quantum logic to perform general calculations. Unlike traditional electronic computers, the objects used by quantum computing to store data are qubits, which use quantum algorithms to operate on data.
‘Quantum Advantage’, often called ‘quantum supremacy’, hopes to use quantum computers to solve problems that are difficult for classical computers to solve. Although the problem itself does not necessarily require practical applications, the superiority of quantum computing brings advantages in using quantum computers to solve practical problems faster than classical computers. Thus, from a computational complexity theory perspective, this generally means that the speedup of quantum computers relative to the best classical algorithms is superpolynomial. The term "quantum computing superiority" was first coined by John Preskill in 2012. But the concept of the advantages of quantum computing, especially for simulating quantum systems, goes back to the quantum computing proposals made by Yuri Manin (1980) and Richard Feynman (1981).
Recently, Google announced that they have made significant progress in the development of a fully functional practical quantum computer and achieved quantum supremacy (superiority), that is, surpassing the current world's number one supercomputer.
Their research shows that their quantum computer named Sycamore can complete the calculations of the world's most advanced computer in 47 years in just 6 seconds! The research was recently published on the preprint site ArXiv.
The paper states that Google’s latest Sycamore quantum processor currently has 70 qubits, while the 2019 version only has 53 qubits. Obviously, the increase in qubits means that the performance of quantum computers can be improved exponentially, making the new processor approximately 241 million times more robust than before.
This latest research marks a milestone moment for quantum computing. With its computing advantages, Google's quantum computer is expected to revolutionize various fields, including artificial intelligence. Solving complex problems at unprecedented speed promises to unlock the next generation of artificial intelligence models and push boundaries that have never been surpassed in many fields.
47 years condensed into one moment
Each qubit can exist in a 0, 1, or superposition state at the same time, so the ability to store and process this level of quantum information is difficult and unmatched by even the fastest classical computers.
In this paper, the Google team stated that quantum computers are expected to perform tasks beyond the capabilities of classical computers. ‘We estimate computational costs based on modified classical methods and demonstrate that our experiments exceed the capabilities of existing classical supercomputers,’ say researchers on the Google team.
Even the Frontier supercomputer in Tennessee (which is currently the fastest supercomputer in the world) cannot touch the porcelain of quantum computers. Of course, the premise is that quantum computers unleash their potential. Because traditional computers operate in a binary code language, they are limited to 0s and 1s, and dual states. Quantum computers go beyond this limit.
However, researchers are still not sure how much it will cost to build Google's quantum computer. But the transformative computing power is unquestionable. According to the Google team, it only took 6.18 seconds for the Frontier supercomputer to match the results of the Google-53 qubit computer. However, it would take the same Frontier 47.2 years to match the computing power provided by Google's latest 70 quantum computer.
major milestones
Many experts in the field consider Google's new quantum computer to be a major advance.
Steve Brierley, CEO of Cambridge quantum company Riverlane, called Google's progress a "major milestone." ‘Quantum supremacy? We don’t need to argue about this issue anymore. ’
Similarly, a professor who is director of the Sussex Center for Quantum Technologies praised Google for solving certain academic problems that are difficult for conventional computers to compute. He emphasized that the critical next step before us is to create a quantum computer that can correct its own inherent operating errors.
Although IBM currently has not commented on Google’s latest quantum computer, it is obvious that Google’s progress in the field of quantum computing has attracted the attention of researchers and companies around the world.
Open up new prospects
There is no doubt that this will open up new prospects and competition for the development of computing technology.
In the study, the Google team mentioned that noise competes with coherent evolution and destroys long-range correlations, making it a huge challenge to fully utilize the computing power of recent quantum processors.
The researchers performed random circuit sampling (RCS) experiments in which they identified different phases driven by the interaction between quantum dynamics and noise.
In quantum computing, this involves testing the performance of a quantum computer by running random circuits and analyzing the resulting output to assess its ability and efficiency to solve complex problems.
Driven by circuit depth, the system first undergoes a dynamic phase transition in which the output distribution is no longer concentrated in one part of the bit string. The second is a transition driven by noise.
Using the cross-entropy benchmark, the researchers observed phase boundaries, which can define the computational complexity of the quantum evolution of noise. At the estimated computational cost of the simulation, 53 qubits completed 1 million noise samples 6.18 seconds faster than a classical computer. And 70 qubits is 47.2 years faster.
Table 1, List of estimated simulation computation times
Finally, the Google team demonstrated a 24-cycle 70-qubit RCS experiment with an estimated fidelity of 1.7-107%, which means that at the same fidelity, the circuit volume increased by about 60%. Google estimated the computational costs based on improved classical methods and demonstrated that the new quantum computer has capabilities beyond existing classical supercomputers.
The Google team said that despite the success of RCS so far, finding practical applications for near-term noisy quantum processors remains an outstanding challenge. The experiments they performed provide a study of how quantum dynamics interact with noise.
The phase boundaries they observed provide quantitative guidance for seven systems in which high-noise quantum devices can properly utilize their computing power. In the weak noise stage, global correlation dominates XEB. This fact protects RCS from spoofing attacks. These are design directions for future applications.
‘Quantum supremacy’ a mistake?
In fact, as early as 2019, Google claimed to have achieved quantum supremacy.
The researchers' paper, published on NASA's website, caused a stir as soon as it was released. The paper at the time stated that Google's processor could perform a calculation in 3 minutes and 20 seconds, while using Summit, the most powerful supercomputer today, to perform the same calculation would take about 10,000 years. Subsequently, Google’s paper was officially published in Nature. So far, the paper has been cited 6225 times!
The paper's corresponding author, John Martinis, and colleagues describe the technological advances that have led to quantum supremacy. They developed a processor composed of 54 qubits (called the Sycamore processor).
The processor exploits quantum superposition and quantum entanglement to achieve an exponential increase in computational space compared to what classical bits can achieve. Since one qubit cannot work effectively, the processor actually uses only 53 qubits.
To test the system, the team designed a task to sample random numbers generated by a quantum circuit. For a classical computer, the difficulty of this task increases with the number of qubits in a quantum circuit. In the end, the quantum processor collected 1 million samples from the quantum circuit in about 200 seconds, which would take today's most powerful supercomputers about 10,000 years to complete. One task. Nature magazine stated that ‘Google’s achievement of quantum supremacy is undoubtedly a remarkable achievement’.
However, criticism and doubts about Google’s “quantum supremacy” incident also followed. The IBM team writes, 'On a classic system, an idealsimulation of the same task could be completed in 2.5 days with much higher fidelity many. ’ This means that Google has not actually demonstrated quantum supremacy, and the competition continues.
Although this result may not look ideal to laymen, it is still of great significance to scientists in the quantum field because it intuitively demonstrates the power of "quantum supremacy" for the first time.
Five years later, traditional computers represented by supercomputers have also experienced explosive development.
In the latest supercomputer ranking list released in June 2023, the first place is the Frontier supercomputer jointly manufactured by AMD and HPE Cray in the United States. Its theoretical maximum throughput reaches 2 exaflops, that is, two quintillion calculations per second. This This means that its computing power is ten times that of the Summit supercomputer.
However, under the impact of dimensionality reduction by quantum computers, "no matter how thick a mosquito's legs are, it cannot defeat an elephant", and the gap between the two has become even wider.
According to the Google research team, it will take 6.18 seconds for the Frontier supercomputer to match the calculation results of Google's -53 qubit computer, and it will take at least 47.2 years for the Frontier supercomputer to match Google's latest quantum computer.
Application prospects in many fields
The following are specific applications of future quantum computers in eight fields:
- In the field of medicine, it may be possible to use quantum computers to accurately predict the mutation direction of viruses based on their transmission routes among different populations and design specific drugs in advance.
- In the field of cryptography, it may take hundreds of thousands of years for a traditional computer to crack the 1024-bit RSA algorithm, but a 512-qubit quantum computer can theoretically crack it in 1 second.
- In the biological field, redefining a species through quantum computers integrating multiple sets of data from genomes, transcriptomes, proteomes, metabolites, and phenotypes will be a focus and difficulty in future research.
- In the field of weather forecasting, the use of quantum computers can greatly improve the real-time calculation of weather forecasts, thereby greatly improving the forecast accuracy and extending the number of forecast days.
- In the field of autonomous driving, cars are equipped with quantum computers that can accurately measure and calculate the elements involved in a safe and comfortable driving scene.
- In the field of auxiliary design, quantum computers connect to the human brain and present the images imagined by the human brain on the screen, and then through simple selections, professional-level designs can be completed.
- In the field of financial management, quantum computers calculate various variables and parameters to predict the direction of a hugely complex system, and can even calculate the probability of a financial crisis to prepare for a rainy day.
- In the field of biological evolution, the use of quantum computers may be able to establish the history of the earth with a time resolution of years or even days.
Of course, there are still many applications for quantum computing. If you want to experience more concrete applications, you might as well give them enough time to grow. We will try them one by one in the future.
references: