Father of quantum computing the technology is far from practical
David Deutsch, the "father of quantum computing" as some call him, when asked if he was surprised that the idea had become a practical technology so quickly, replied with characteristic sternness: "Not yet (practical)".
How well do quantum devices really perform in solving real-world problems? Recently, Garnet Chan's team at Caltech answered this question by simulating molecules and materials using a 53-quantum-bit Google processor called Weber, based on Sycamore. The results show that quantum circuits are still unable to outperform classical circuits when it comes to simulating molecules.
The Google-made Sycamore chip contains 53 quantum bits and is often used to explore the "superiority of quantum computing".
01Quantum chip performs poorly when simulating molecular materials: error correction, noise reduction urgently needed
"Our goal is to understand how well the Sycamore hardware performs in a class of physically relevant circuits with physically relevant success metrics." Garnet Chan's team at Caltech used a 53-quantum bit chip associated with Google's Sycamore to simulate real molecules and materials. They chose test cases, one of which is an eight-atom cluster of iron and sulphur (FeS) in the catalytic core of a nitrogenase enzyme - which fixes atmospheric nitrogen into a bio-available form, a process whose understanding is valuable for the development of artificial nitrogen fixation catalysts - and another is the crystalline material alpha ruthenium trichloride (alpha-RuCl3), which is a compound of great interest in the field of quantum materials, as it is thought to display the peculiar low-temperature phase of spin liquids [1].
Illustration of the core steps of catalysis. In the catalytic site of nitrogen fixation enzymes, which are responsible for extracting nitrogen from the atmosphere (nitrogen fixation), there are clusters of iron (red) and sulphur (yellow) atoms that catalyse the splitting of nitrogen molecules. The researchers want to simulate this process on a quantum computer in order to develop artificial nitrogen fixation catalysts.
How well does the chip perform? Frankly, fairly average, and Chan admits that he initially thought that with 53 quantum bits at their disposal, they would be able to simulate these systems without any problems; however, as he learned more about the problem, he was disabused of this idea. By mapping them onto quantum circuits, the researchers could reasonably try to calculate them; however, despite the advantages of the energy spectrum of FeS clusters and the heat capacity of α-RuCl3, the classical method performed at least as well as the results.
One of the main obstacles to accurate quantum simulations is noise - random errors in the switching of 'gates' performing quantum logic operations and in the reading of output states. These errors accumulate and limit the number of gate operations that can be performed before the noise dominates the calculation. The researchers found that simulations with more than 300 gates were swamped by noise; but the more complex the system, the more gates were needed. Fe-S clusters, for example, have long-range interactions between spins; many gates are needed for such interactions to be accurately represented.
Due to these challenges, on-chip simulations are quite limited. For example, the simulations provide a fairly good prediction of the energy spectrum and heat capacity of Fe-S clusters, α-RuCl3: but only if the simulation system is not too large and the team can only obtain meaningful results for very small 6-atom lattice blocks; if they increase the size to just 10 atoms, the noise swamps the output. And restrictions on gate manipulation meant that only about one-fifth of Weber's quantum resources were available for computation. However, when Chan and colleagues turned to simulating model systems more suited to Weber's specific circuit architecture, they could increase this usage to half.
As a result, Chan says that unless there are better ways to reduce noise or correct errors, it is difficult to see how quantum circuits will perform better at solving such problems.
02Is quantum computing really useful?
In a lecture at MIT in 1981, Feynman talked about "simulating physics with computers". Although it was already being done at the time, Feynman said he wanted to talk about "an exact simulation that computers would do exactly as nature does". At the same time, he pointed out that since nature is quantum-mechanical, what was needed for this was a quantum computer.
But history is still in the making. David Deutsch, the visionary physicist who raised the possibility of quantum computing in 1985, when asked if he was surprised that the idea had become a practical technology so quickly, replied with characteristic sternness: "Not (yet)".
Deutsch's reluctance to accept that practical quantum computing has arrived presumably stems from the question of whether it can do anything really useful. Of course, one could construct a problem that is very difficult for classical devices, but ideal for quantum computers, and then demonstrate that just a few dozen quantum bits are sufficient to achieve "quantum computing superiority".
But how helpful is this, as we all know, in the real world?
When Feynman described the idea of quantum computing, he had in mind that such a facility would be used to simulate systems governed by quantum laws, such as molecules and materials. Instead of using the cumbersome classical approximation, the standard ab initio approach to quantum chemistry, quantum state representations of atoms and molecules would be used to calculate properties such as energy spectra, electronic band structures and stability.
Quantum computers have been doing this for several years. Alán Aspuru-Guzik and his colleagues at the University of California, Berkeley, showed back in 2005[2] that it was possible to simulate simple molecules such as water and lithium hydride using just a few quantum bits. And in 2017, an IBM team used a simple six-quantum-bit circuit to simulate LiH and BeH2 - the latter being the first triatomic substance to be simulated in this way [3].
03Keeping it sane: breakthroughs in technology are still to come
Given the growth in available resources over the past few years, one might think we could do more now; but the results of this experiment have once again pushed attitudes towards quantum simulations towards sanity.
"These results are state-of-the-art and they show the challenges that will need to be overcome in terms of the performance of future devices," says Alán Aspuru-Guzik[4] of the University of Toronto, an expert in the use of quantum computing in chemistry and materials. "These results are both exciting and daunting, and they show that we still have a lot of work to do."
Reference links:
[1]https://physics.aps.org/articles/v15/175
[2]https://www.science.org/doi/10.1126/science.1113479
[3]https://www.science.org/doi/10.1126/science.1113479
[4]https://www.chemistryworld.com/opinion/quantum-computing-has-its-limits/4016553.article
