Quantum computers may not be able to simulate molecules and solve other problems in chemistry as fast as researchers had hoped, a new analysis has found.
The promise of quantum computers is that they can solve some tasks exponentially faster than ordinary computers, but this hasn鈥檛 yet been demonstrated for useful, real-world problems.
One highly anticipated application is calculating the energy levels of molecules, an important task in chemistry. But at the California Institute of Technology and his colleagues, including researchers at Google and Amazon, have now examined whether quantum computers really could deliver an exponential advantage, and found that the evidence is lacking.
Advertisement
The team calculated how long it would take for classical and quantum methods to find the quantum states of iron-sulphur groups, which are molecules found throughout nature and frequent targets for quantum chemistry.
A common approach to find a molecule鈥檚 quantum state is to start with a guess and to then fine-tune this until a solution is reached. But Chan and his team found that just determining the initial guess could require so much computing power that the exponential advantage from quantum computing becomes diminished.
They also found that, when compared with quantum methods, there was no indication that classical methods couldn鈥檛 be optimised or tweaked to remove the exponential advantage as well.
鈥淚t鈥檚 bringing a dose of reality into some of the hyperbole that has gone around this particular question,鈥� says at the University of Oxford. 鈥淚t鈥檚 good to do that because it鈥檚 really important to know where we stand scientifically and what we need to do to make progress.鈥�
Though Chan and his team examined the evidence for exponential speed-up, they left open the possibility of a smaller improvement called polynomial speed-up, which could still offer a large advantage over traditional methods. They also acknowledged that quantum computing is still at an early stage, with the potential for new technologies. 鈥淓ven if the benefit isn鈥檛 exponential, if it鈥檚 a significant polynomial benefit, then it can be massively worthwhile,鈥� says Tew.
While the work shows that commonly used quantum algorithms might not provide an exponential speed-up, there is also the possibility that specific systems and future algorithms might. 鈥淚t鈥檚 very hard to be exhaustive when it comes to the full range of chemical systems that can be simulated with classical methods,鈥� says at Lancaster University, UK. 鈥淪o there鈥檚 potentially areas that they haven鈥檛 considered where quantum computing would offer this exponential advantage.鈥�
Because there are still large advantages for quantum computers in solving chemistry problems, the results are unlikely to change or halt any current quantum computing research, says Tew. 鈥淗opefully, it will bring more clarity in the way that questions are being asked and answered.鈥�
Reference:
