A record-breaking milestone has been reached in quantum computing, which could mean that unthinkably fast processing is now a realistic prospect.
Scientists at Sussex University have managed to transfer data between chips at record speeds and - more importantly - record accuracy.
"What we have achieved here is the ability to realise extremely powerful quantum computers capable of solving some of the most important problems for industries and society," said lead researcher Prof Winfried Hensinger.
Speed and accuracy
Quantum computing relies on a few principles in quantum psychics, namely that subatomic particles can be in two places at once, and can be mirror each other's actions almost simultaneously across unfathomable distances.
These properties mean that computers could potentially handle multiple processes at speeds not possible with even today's best computers. They have been in development for over two decades now, but so far, only small systems that have limited uses are currently up and running. Big tech companies such as IBM, Google and Microsoft have some of their own machines.
One of the major obstacles for the technology's development has been the ability to transfer information across chips so it remains intact. To make quantum computers work, they are by design highly sensitive and so have a low fault tolerance. That means that the slightest interferences can disrupt their effective operation.
But publishing results in the peer-reviewed journal Nature Communications, the research team at Sussex University demonstrated a way to transfer information between quantum chips with 99.999993% reliability, the connection rate was 2424/s.
Both set world records, the researchers say, and show it's possible for quantum chips to be fitted together to build more powerful quantum computers.
Director of the National Quantum Computing Centre, Prof Michael Cuthbert, commented on the findings:
"To build the type of quantum computer you need in the future, you start off by connecting chips that are the size of your thumbnail until you get something the size of a dinner plate. The Sussex group has shown you can have the stability and speed for that step."
However, he added, "You need a mechanism to connect these dinner plates together to scale up a machine, potentially as large as a football pitch, in order to carry out realistic and useful computations, and the technology for communications for that scale is not yet available."
If quantum computing does become practically applicable, it could mean big things for all sorts of industries. It could lead to new discoveries in science, since they can perform calculations that are impossible for any human or current computer to achieve.
Substantial improvements in AI are also possible with quantum computers. Currently, it can take months to train an AI model to become effective. Not needing to rely on linear binary of standard computers, where information is processed as either a 1 or 0, quantum computing can hold two states of information at the same time, speeding up its operation considerably.
In fact, IBM has already revealed a mathematical proof that quantum machine learning is exponentially faster than standard methods of ML, as long as "one can provide classical data to the algorithm in the form of quantum states." Although it remains theoretical at this point, if it can be applied then the future for AI and quantum computing looks promising.
- Best processors: For top CPU performance
