Tech news: Enter quantum supremacy!
It is almost needless to say that this is a very important achievement with regard to the future of quantum computing. Until now quantum computers could only solve problems faster than classical computers.
One of the biggest challenges of extremely powerful quantum computers is that they are more susceptible to errors than classical computers due to noise and decoherence (the loss of information from a system into the environment as a result of entanglement between the system and environment).
Various quantum error-correcting codes are used to counter this problem, but until now it was unclear if these correction methods were scalable with the increase in the number of qubits.
But let us first take a step back to explain quantum computing. Quantum computers are fundamentally different. Traditional computers use the flow of electricity that is switched on and off inside the circuits to generate zeros and ones (bits). Simple rules to turn these switches on and off can be used to solve any mathematical problem.
Quantum computers, on the other hand, are not built upon using the flow of electricity, but rather rely on the physical properties of electrons, photons and other small bits of matter subject to the laws of quantum physics.
The states of tiny bits of matter can best be described as amplitudes (although the bits of matter can act as particles and waves). A particle can have two amplitudes at the same time, a state that is called superposition. They can also be entangled since a change in one particle can instantly change another particle.
Furthermore, particle amplitudes can cancel out one another (interference), but most interestingly, the particles do not really exist as a point in space, but exist as a probability of existing.
Unlike classic computers that use bits (ones or zeros), quantum computers uses quantum bits (qubits) as the basic units of information. While the binary approach of classical computers store data as either a one or zero, quantum computers take advantage of the quantum phenomenon known as superposition where qubits can be ones and zeros simultaneously. This increases the amount of information that can be processed exponentially.
Google’s breakthrough was announced in a paper by several authors in the highly esteemed Nature Journal of October 23 (volume 574:505-510) under the title Quantum supremacy using a programmable superconducting processor.
For this experiment, Google has partnered with the University of California to build the hardware for the quantum computer and also partnered with Nasa to help test it.
In the paper, the researchers state that according to their knowledge, the quantum computing experiment was the first computation that can only be performed by a quantum processor.
The experiment involved the proving of the randomness of numbers generated by a random number generator (random circuit sampling). While it took Sycamore only 200 seconds to repeat the sampling process a million times, it would have taken the world’s fastest traditional supercomputer, Summit, about 10000 years to achieve the same full computation result, effectively proving quantum supremacy.
Google used a 54 superconducting-quantum-bit computer system called Sycamore to achieve this important breakthrough by parallel processing 253 states, which is 10 million billion states (the Hilbert space of dimension). Although one of the qubits did not function, the remaining 53 were entangled with one another and were used to produce a set of binary digits and check that their distribution was truly random.
To demonstrate quantum supremacy, Google compared a quantum computer simultaneously with a classic computer and gradually increased the complexity of the problem until the classic computer was not able to keep up.
With this quantum computing achievement that cannot realistically be imitated on a classical computer, Google has opened up a new area of computing that could change our future and the solving of complex problems.
However, some scientists are very sceptical and point out that quantum computers capable of solving practical tasks are still years away. IBM’s head of research, Dario Gil, argues that Google’s system is a specialised piece of hardware that was designed to solve a single problem and that it was not a general-purpose computer.
In a paper posted on arXiv.org, researchers from IBM further suggested that the calculation that Google claims would take a traditional computer 10000 years to solve, could in fact be solved in only 2.5days using an improved technique on an extremely powerful supercomputer.
They are therefore of the opinion that Google cannot really claim quantum supremacy, despite the fact that the quantum computer would still be significantly faster than the supercomputer (200 seconds versus 2.5 days). And the quantum computer does this with only one small chip while the Summit supercomputer of 520.3m² could fill two-and-a-half tennis courts.
Quantum supremacy may be rather nebulous and not as straightforward as some researchers would like to suggest, but what was really remarkable is the new level of control over famously unstable qubits.
Algorithms on classical computers will no doubt continue to improve and possibly outperform the particular quantum computer used by Google. However, quantum computers are improving exponentially and much faster.
Professor Louis C H Fourie is a technology strategist. [email protected]