Special relativity could be harnessed to build a novel quantum computer, and creating it this way could let us use machine learning to deepen our understanding of the quantum realm.
Albert Einstein鈥檚 theory of special relativity describes how moving at close to the speed of light would affect travellers鈥� experience of space and time. These insights don鈥檛 merely give us thought experiments; they are crucial for technologies such as satellite communication and GPS.
Now, at the Perimeter Institute for Theoretical Physics in Canada and his colleagues say such relativistic effects could help develop a particular kind of quantum computer. And they constructed the most complete and detailed mathematical model yet for building one.
Advertisement
Previously, researchers have explored how some quantum computing operations could arise from relativistic effects. For example, two quantum bits, or qubits, can become connected through quantum entanglement when they move next to each other extremely quickly. But until now, there was no 鈥渞ecipe鈥� for how to combine all of those operations into a full mathematical model of a computer.
鈥淚t was like they said 鈥楬ey, we have the flour, we have the water, we have the yeast, clearly it鈥檚 possible to make bread.鈥� And we really detailed how much yeast, how much flour, what鈥檚 the order, how much water,鈥� says Perche.
He and his colleagues turned to the mathematics of relativistic quantum information theory, which explains how ultra-fast motion and interactions can edit the information encoded in qubits. It also takes into account the effects of interactions between qubits and the quantum fields that permeate all space.
The researchers also used machine learning and wrote short quantum computing programs themselves. Then they used an algorithm to find how the quantum computer鈥檚 qubits would have to move to run those programs.
The researchers tested their method by determining how their relativistic quantum computer could run the quantum Fourier transform algorithm. This is a relatively simple but ubiquitous program that all quantum computing researchers would find useful, says at the University of Innsbruck in Austria, who worked on the project. For instance, it is part of Shor鈥檚 algorithm, a quantum program with the potential to crack encryption by factoring large numbers.
at the University of Waterloo in Canada, who wasn鈥檛 part of the team, says the new work is a significant leap forward for relativity-based quantum computers. 鈥淭he authors carefully and systematically demonstrate that such a scheme is indeed viable,鈥� he says.
Past studies have suggested that some versions of relativistic quantum computers could actually be built with qubits made from tiny superconducting circuits. LeMaitre says his team鈥檚 scheme is not necessarily incompatible with how quantum computers are built today 鈥� but there are years worth of theoretical and engineering details that would have to be ironed out before a relativistic quantum computer can be made.
Even before that, Martin-Martinez says the model could lead to new understandings of both quantum information processing and quantum fields 鈥� and offer a deep way to connect the two.
Because machine learning is already part of their approach, the researchers want to explore how it could be used to learn new properties of quantum fields or perhaps even uncover new laws of physics that those fields must follow.
Journal reference:
Physical Review Letters,
