Quantum computers promise to leave our current computers behind, coughing and spluttering in the dust, but not all computers created with quantum technology can keep this promise.
This is why the international scientific community, besides promoting research on quantum computing, has also addressed the issue of benchmarking quantum computing.
A fundamental contribution to this issue has been provided by the extensive research projects conducted by two teams and that has recently been published on Nature Photonics.
The first research team is coordinated by Sapienza Physicist Fabio Sciarrino in collaboration with the CNR Institute for Photonics and Nano-Technology and the Universidad Federal Fluminense in Brazil, while the second is directed by Anthony Laing at the University of Bristol.
Starting with the hypothesis that a traditional computer cannot keep up with the calculations of a quantum computer, the researchers decided to exclude traditional tools to perform a specific test operation. Thus, both groups decided to use a simple quantum simulator with Boson Sampling. In 2010, two MIT researchers (Aaronson and Arkhipov) had theoretically demonstrated that the most powerful traditional computers could not solve this algorithm.
The two teams developed simple quantum systems that could correctly solve Boson Sampling in order to devise tools and techniques that could also certify more complex quantum computers.
The Sapienza team employed the most advanced integrated photonics applications. The experiment was to study the behaviour of photons (light quants) within a complex structure, a complex network developed on a glass chip by a laser. Three undistinguishable light photons were injected into the glass chip and propagated on increasingly larger interferometers (5,7,9 and 13 paths); indeed, the greater the path complexity, the greater the complexity of the simulated quantum computer.
According to Prof. Sciarrino, “not only can Boson Sampling not be reproduced by conventional technology, but its results cannot be verified by a traditional computer or any other known method. We demonstrated how to measure certain properties of Boson Sampling and how to provide experimental proof of its correct functioning. Our experimental results open a new road for the certification of complex quantum systems.”
The Italo-Brazilian research project was part of Project ERC 3D-QUEST, which aims to demonstrate the potential of linear optics for the implementation of computing more powerful than that provided by current computers.
The Bristol team, on the other hand, used two integrated optical circuits implanted in different silicon-based materials and injected up to 5 photons into them. Bristol Researcher Anthony Laing pointed out "with only three or four more photons, our experiments would have become so complex that we could not have monitored them on our laptops. In any case, the testing techniques that we have developed, should soon be able to find a practical application.”