BEIJING, Jan. 7 ( Xinhua) — According tomedia reports, Google announced in the fall of 2019 that its quantum computers are computing much faster than the top supercomputers, showing that “quantum supremacy” has been achieved, and IBM was quick to object. Calling your classic supercomputer not only computing at the same speed as Google’s quantum computer, but also killing Google in seconds of authenticity, one should look at Google’s announcement with a “skeptical eye”.
It is not the first time quantum computing has been questioned. Last year, Michel Dyakonov, a theoretical physicist at the University of Montpellier in France, published an article on IEEE Spectrum, the flagship journal of electronicand and computer engineering. Why we’ll never be able to make a practical quantum supercomputer” lists a number of reasons. Subhash Kak, an expert on quantum computing at Oklahoma State University, agrees that it is difficult to create truly useful quantum computers because random errors in hardware are unavoidable.
What is a quantum computer?
To understand why, it is important to understand how quantum computers work, because their principles are fundamentally different from that of classical computers.
Classic computers use countless 0s and 1s to store data that can represent voltages at different points on a circuit, but quantum computers use qubits that can be thought of as a series of waves with amplitudes and phases.
The nature of qubits is so special that they can exist in superposition states, i.e. at the same time they can be either 0 or 1, and qubits can be entangled with each other and share the same physical properties even when they are far apart. This behavior does not exist in the classical physics world, and as soon as the experimenter tries to interact with the quantum state, the superposition disappears.
Because of the superposition state, a quantum computer with 100 qubits can give 2,100 solutions at the same time. This exponential-level parallel computing undoubtedly has a huge speed advantage when solving specific problems, such as code cracking classes.
There is also another quantum calculation method, called Quantum Anne, which uses quantum bit acceleration to solve optimization class problems. Canada’s D-Wave Systems has built a range of optimization systems using qubits, but critics say they perform no better than classic computers.
Still, companies and national governments are investing heavily in quantum computing. The European Union has a $1.1 billion master plan for quantum projects, and the U.S. National Quantum Initiative Provides $1.2 Billion to advance quantum information science over a five-year period.
Cracking encryption algorithms is a powerful motivation for many countries to study quantum technology, and if they succeed in mastering it, they will gain a huge intelligence advantage, in addition to these investments, these investments have given a strong impetus to the study of basic physics.
Many companies are doing their best to build quantum computers, including Intel, Microsoft, IBM and more. These companies are developing hardware to simulate classic computer circuit models. However, there are fewer than 100 qubits in today’s experimental system, and to truly have computing power, computers must have hundreds of thousands of qubits to do so.
Google’s Sycamore chip needs to be placed in a cryogenic thermostat to keep it cool.
Noise and error correction
The mathematical principles behind quantum algorithms are clear, but there are still huge challenges in technology.
If a computer wants to function properly, it must be able to correct random small errors at any time. In quantum computers, these errors may come from problematic circuit elements, or the interaction between qubits and their surroundings. Once these problems arise, the interconnectedness between qubits quickly disappears, so the calculation time must be shorter than this time, and if these random errors are not corrected, the calculations of quantum computers are worthless.
In classic computers, small-scale noise can be corrected using the so-called “threshold” concept, similar to the rounding of numbers. In the case of an integer transfer, assuming that the known error value is less than 0.5, if the number received is 3.45, it will be automatically corrected to 3.
More serious noise can be corrected by introducing “redundancy”. Suppose 0 and 1 are transmitted in the form of 000 and 111, and only one bit in the transmission process will go wrong, so that if the number received is 001, it will be automatically corrected to 0, and if 101 is received, it will be corrected to 1.
Quantum error-correcting code is a generic version of the classic computer error correction code, but there is a key difference between the two. First, unknown qubits cannot be copied, so redundant error correction cannot be applied. Second, errors in the data entered before error correction code was introduced cannot be corrected.
Although noise is a major challenge for quantum computers, this is not the case with quantum encryption. Because in quantum encryption, there is no interstitity between the individual bits, and a single qubit can be isolated from the outside environment for a long time. Quantum encryption allows two users to exchange so-called “keys” (usually a long string of numbers), which is like a key to protect data, and no one in the key exchange system can crack it. This type of key exchange system can be used for encrypted communication between satellites and naval vessels. However, the true encryption algorithm used after exchanging keys is still a classic algorithm, so in theory the level of encryption is not higher than that of the classic encryption method.
Quantum encryption has been used in a few large banking transactions, but since both parties must be authenticated through a classic protocol, which is the weakest link in the chain, the strength of the entire encryption system is not much different from the existing system. Banks are still using an authentication process based on classical encryption methods that can be used for key exchange in itself without losing the overall security of the system.
Therefore, if quantum cryptography is to be more secure than the existing technology, it must shift the focus to quantum information transmission.
Challenges in quantum computing at commercial scale
If the problem of quantum information transmission can be solved, quantum encryption technology is still promising, but quantum computing is not necessarily. Error-correcting is already so important to ordinary multi-function computers, and a huge challenge for quantum computers, so it’s hard to build a commercial-scale quantum computer. (Leaf)