A quantum computer has been used to design an improved qubit that could power the next generation of smaller, higher-performance and more reliable quantum computers. Exploiting the ability of quantum processors to simulate the behaviour of quantum circuits that classical computers can’t could let us quickly develop prototypes.
As classical computer chips became more complex and grew from having dozens of components to thousands, millions and even billions, it quickly became impractical to design them manually. For decades, it has been commonplace to use computers themselves to help create and optimise new chip designs for the next generation of computers.
But it is unfeasible to simulate the operation of all but the simplest quantum processor inside a classical computer. This is because the computing resources required grow exponentially as each new qubit is added.
Advertisement
Now researchers at the University of Science and Technology of China in Shanghai have applied the same bootstrapping approach used in classical computing to design new quantum computers. The research has led to the invention of a new type of qubit called a plasonium, which is physically smaller, less noisy and also able to retain its state longer than the group’s current qubit design.
The team believes that the work opens the way to designing advanced quantum processors using existing machines.
at Imperial College London says the concept is intriguing but also intuitive. “If you think about a classical machine that’s trying to simulate a quantum processor, it’s hard work. So it’s obvious that once you’ve got a quantum processor you’d see if you can use it for that purpose. Already it’s given them some new insights.”
Knight says that the new qubit design offers advantages but also, crucially, reduces several inconvenient features of current generation circuits.
An algorithm called a variational quantum eigensolver, which is often used in quantum chemistry to calculate the energy levels of molecules, allowed the team to simulate the behaviour of particles in quantum circuits and iron out negative properties while developing positive features, all without having to build vast numbers of physical prototypes.
The plasonium qubits are each only 240 micrometres long, which is just 40 per cent of a typical qubit called a transmon, and using them will allow today’s large processors to be miniaturised. This will be key because future quantum computers will need to scale up from carrying dozens of qubits, as most do today, to carrying millions or billions to carry out useful tasks.
Reducing “noise” in quantum computers is another vital hurdle between current machines and powerful, useful quantum computers. Crucially, the plasonium is also less noisy than the group’s existing transmon qubit. Error-correction is a standard feature of classical computers to combat occasional noisy “bit flips” where a single bit encounters an error due to a charged particle racing through the universe and colliding with an electron inside a chip. But the problem is far more complex and more difficult to solve in quantum computing.
The new type of qubit also displays another desirable feature called strong anharmonicity. A qubit will have two possible states, 1 and 0, but often there will be other states that can also be reached accidentally, causing computational errors. If these undesirable additional states lay at regular intervals of power after 0 and 1 then the chance of accidentally falling into them during operation is higher. The plasonium’s additional states are varied and therefore less likely to be found accidentally. This means the plasmodium will experience fewer problems with calculations during operation.
Reference:
