In 20 19, Google used a device with 54 qubits (the quantum equivalent of conventional computing bits) for the first time to perform a basically useless calculation called random sampling calculation, thus achieving this goal. In 20021year, a team of the University of Science and Technology of China solved a more complicated sampling problem with 56 qubits, and later pushed it further with 60 qubits.
But Bob Sutor of IBM said that this leap-forward game is an academic achievement and has not yet had a real impact. Only when the quantum computer is obviously superior to the classical computer and can solve different problems can it achieve real hegemony, instead of the current random sampling calculation as a benchmark.
He said that IBM is striving to achieve the "quantum business advantage"-in this respect, quantum computers can solve really useful problems for researchers or companies faster than traditional computers. Sutor said that this has not come yet, and the new year will not come, but it can be expected within ten years.
Nir Minerbi, co-founder of quantum software company Classiq, is more optimistic. He believes that the new year will show quantum hegemony on a useful issue.
Remember when the first electric car came out? They are useful for driving to the grocery store, but they may not be suitable for driving 300 kilometers to send children to college. Just like electric cars, quantum computers will get better and better over time, making them play a role in a wider range of applications.
There are many obstacles to solving practical problems. The first is that devices need thousands of qubits to do this, and these qubits must be more stable and reliable than the existing ones. Researchers are likely to need to combine them to work as a single "logical qubit". This helps to improve fidelity, but weakens the scale: thousands of logical qubits may require millions of physical qubits.
With the passage of time, quantum computers will become better and more useful in a series of applications.
Researchers are also committed to quantum error correction in order to repair faults when they occur. In July, 20021,Google announced that its Sycamore processor could detect and repair errors in its superconducting qubits, but the extra hardware needed to perform this operation introduced more errors than repaired them. Researchers at the Maryland Joint Quantum Research Institute later successfully passed this critical break-even threshold with their captured ion qubits.
Even so, it is still too early. If the general quantum computer solves a useful problem in the new year, it will be "quite shocking". Protect a single coded qubit at any time, let alone calculate thousands of coded qubits.
How big does a quantum computer need to crack bitcoin encryption or imitate molecules?
It is expected that quantum computers will be subversive and may affect many industrial fields. Therefore, British and Dutch researchers decided to explore two completely different quantum problems: cracking the encryption of Bitcoin (a kind of digital currency) and simulating the molecules responsible for biological nitrogen fixation. The researchers described a tool they created to determine how big a quantum computer is needed to solve such problems and how long it will take.
Most of the existing work in this field focuses on specific hardware platforms and superconducting devices, which are also being developed by IBM and Google. Different hardware platforms will have great differences in key hardware specifications, such as the operation speed and control quality of qubits. Many of the most promising quantum advantage use cases will require error-correcting quantum computers. Error correction can run a longer algorithm by compensating the inherent errors in quantum computers, but at the cost of more physical qubits. Extracting nitrogen from the air to produce ammonia used as fertilizer is very energy-consuming, and improving this process may affect the world food shortage and climate crisis. At present, the simulation of related molecules even exceeds the ability of the fastest supercomputer in the world, but it should be within the scope of the next generation of quantum computers.
Our tool automatically calculates the error correction overhead according to the key hardware specifications. In order to make the quantum algorithm run faster, we can perform more operations in parallel by adding more physical qubits. We introduce additional qubits as needed to achieve the required running time, which largely depends on the running speed of physical hardware. Most quantum computing hardware platforms are limited, because only adjacent qubits can interact directly. In other platforms, such as the design of some trapped ions, qubits are not in a fixed position, but can be physically moved-which means that each qubit can directly interact with a large number of other qubits.
We explored how to make full use of this ability to connect distant qubits, so as to solve the problem in a shorter time with fewer qubits. We must continue to adjust the error correction strategy to make use of the underlying hardware, which may enable us to use a smaller quantum computer than previously assumed to solve far-reaching problems.
Quantum computers are more powerful than classical computers in cracking many encryption technologies. Most secure communication devices in the world use RSA encryption. One of RSA encryption and bitcoin (elliptic curve digital signature algorithm) will be vulnerable to quantum computing one day, but today, even the largest supercomputer will never pose a serious threat. Researchers estimate that the quantum computer needs to be large enough to crack the encryption of the bitcoin network in a short time that actually poses a threat-before it is announced and integrated into the blockchain. The higher the transaction cost, the shorter the window, but it may range from a few minutes to several hours.
The most advanced quantum computer has only 50- 100 qubits. "We estimate that we need 30 million to 300 million physical qubits, which shows that Bitcoin should be considered safe at present and will not be attacked by quantum, but devices of this size are usually considered achievable, and future progress may further reduce the requirements. Bitcoin network can' hard fork' quantum security encryption technology, but this may lead to network expansion problems due to increased memory requirements.
The researchers emphasized the improvement speed of quantum algorithm and error correction protocol. Four years ago, we estimated that ion trap equipment needed 1 100 million physical qubits to crack RSA encryption, which required a device with an area of 100 x 100 square meter. Now, with the overall improvement, this may be greatly reduced to only 2.5 x 2.5 square meters. Large-scale error correction quantum computer should be able to solve important problems that classical computers can't solve. Analog molecules can be applied to the development of energy efficiency, batteries, improved catalysts, new materials and new drugs. Further applications exist in all directions-including finance, big data analysis, fluid flow, and logistics optimization of aircraft design.
What is the quantum apocalypse?
Imagine a world where encrypted secret files are suddenly cracked-this is the so-called "quantum revelation". In short, quantum computers work in a completely different way from those developed in the last century. In theory, they may eventually be many, many times faster than today's machines. This means facing an extremely complex and time-consuming problem-such as trying to decrypt data-in which there are billions of permutations, and if there are any, it will take an ordinary computer many years to crack these passwords. But in theory, the future quantum computer can complete this work in a few seconds. Such a computer can solve various problems for human beings. The British government is investing in the National Quantum Computing Center in Havel, Oxfordshire, hoping to completely change the research in this field.
A New Language for Quantum Computing
Twist is a programming language developed by MIT, which can describe and verify which data are intertwined to prevent quantum programs from making mistakes. Time crystallization, microwave oven and diamond, what are the similarities between these three different things? Quantum computing. Unlike traditional computers that use bits, quantum computers use qubits to encode information as 0 or 1, or both. Coupled with various forces from quantum physics, these refrigerator-sized machines can handle a lot of information-but they are far from perfect. Just like our ordinary computers, we need the correct programming language to calculate correctly on quantum computers.
To program a quantum computer, we need to know something called entanglement. Entanglement is a computer used for all kinds of qubits, which can be converted into powerful energy. When two qubits are entangled, the effect on one qubit can change the value of the other qubit, even if they are physically separated, which leads to Einstein's description of "long-distance ghost interaction". But this effectiveness is also the source of weakness. When programming, discarding one qubit without paying attention to its entanglement with another qubit will destroy the data stored in the other qubit, thus endangering the correctness of the program.
Scientists in Computer Science and Artificial Intelligence (CSAIL) at MIT aim to solve this mystery by creating their own quantum computing programming language Twist. Twist can describe and verify what data is entangled in quantum programs in a language that classical programmers can understand. This language uses a concept called purity, which does not force entanglement, produces a more intuitive program, and ideally has fewer errors. For example, programmers can use Twist to indicate that the temporary data generated by the program will not be entangled with the answers of the program as garbage, so they can be safely discarded.
Although emerging fields may make people feel a little flashy and futuristic, and images of huge metal machines will emerge in their minds, quantum computers may achieve computational breakthroughs in tasks that cannot be solved by classics, such as cryptography and communication protocols, search and computational physics and chemistry. One of the main challenges of computational science is to deal with the complexity of the problem and the amount of calculation required. Classical digital computers need a very large number of exponential bits to deal with this simulation, while quantum computers may use a very small number of qubits to complete this task-if there is a correct program. "Twist, our language, allows developers to write more secure quantum programs by explicitly stating when not to entangle with another qubit," said Charles Yuan, a doctoral student in electrical engineering and computer science at MIT and the first author of a new paper on Twist. "Because understanding quantum programs requires understanding entanglement, we hope Twist can pave the way for developing languages and make it easier for programmers to cope with the unique challenges of quantum computing."
Unlock quantum entanglement
Imagine a wooden box with 1000 cables sticking out from one side. You can pull any cable out of the box or push it in completely.
After doing this for a while, the cables will form a bit pattern-0 and1-depending on whether they are inside or outside. This box represents the memory of a classic computer. A computer program is a series of instructions about when and how to pull a cable.
Now imagine that the second box has the same appearance. This time, when you pull a cable and see it appear, several other cables are pulled back inside. Obviously, in the box, these cables are intertwined somehow.
The second box is the analogy of quantum computer. To understand the meaning of quantum program, we need to understand the entanglement in its data. But detecting entanglement is not simple. You can't see the wooden box. The best you can do is try to pull the cable and carefully reason which ones are wound. Similarly, today's quantum programmers have to reason and struggle by hand. This is a Twist design, which helps to massage some staggered parts.
Twist designed by scientists is expressive enough to write programs for famous quantum algorithms and identify errors in their implementation. In order to evaluate the design of Twist, they modified the program, introduced some errors that human programmers are relatively unaware of, and showed that Twist can automatically identify errors and reject the program.
They also measured the actual execution of the program at runtime, and compared with the existing quantum programming technology, its overhead is less than 4%.
For those who are worried about quantum's "dirty" reputation in cracking encryption systems, Yuan said that it is not clear to what extent quantum computers can fulfill their performance commitments in practice. "Post-quantum cryptography is undergoing a lot of research, and it exists because even quantum computing is not omnipotent. So far, there are a set of very specific applications, in which people have developed algorithms and technologies that quantum computers can surpass classical computers. "
The next important step is to create a more advanced quantum programming language using Twist. Most of today's quantum programming languages are still similar to assembly languages, which string together low-level operations, paying no attention to data types and functions, and also paying no attention to the typical contents in classical software engineering.
Quantum computers are prone to errors and difficult to program. By introducing and inferring the "purity" of program code, Twist has taken a big step towards simplifying quantum programming, ensuring that the qubits in pure code will not be changed by bits that do not exist in the code. This work was partially supported by MIT -IBM Watson Artificial Intelligence Laboratory, National Science Foundation and Naval Research Office.
Notes. quantum computer
Quantum computer is a kind of computing equipment that directly uses quantum mechanical phenomena (such as superposition and entanglement) to calculate data. The basic principle behind quantum computing is that quantum properties can be used to represent data and perform operations on it.
Although quantum computing is still in the primary stage, some experiments have been carried out, in which quantum computing operations are carried out on a very small number of quantum bits (quantum binary numbers). Practical and theoretical research continues, and many governments and military funding agencies support quantum computing research to develop quantum computers for civil and national security purposes, such as cryptanalysis.
If we can build a large-scale quantum computer, we can solve some problems faster than any classical computer (such as Shor algorithm). Quantum computers are different from DNA computers and traditional computers based on transistors. Some computing architectures (such as optical computers) may use classical electromagnetic wave superposition. Without some specific quantum mechanical resources, such as entanglement, guessing cannot surpass the exponential advantage of classical computers.