Researchers at Paderborn University and Prague have investigated a key question in photonic quantum computing: whether quantum sampling experiments remain difficult to replicate on classical computers, even when real experimental noise factors are taken into account. Their findings suggest that the underlying computational complexity remains largely intact despite a multitude of such imperfections – an encouraging sign for efforts to demonstrate a useful quantum advantage using light-based devices.
The work relates to boson-sampling experiments, in which quantum light is guided through complex optical circuits and detected at the output. It is assumed that predicting the detailed pattern of the measurement results will become exceptionally difficult for conventional computers as the scale of such experiments increases. Although these tasks were originally developed to test the limits of quantum computing, they could also find applications in chemistry and in solving certain graphene-related problems.
In practice, however, no laboratory system is perfect. Photons can be lost, equipment can introduce noise, and photons may not behave as identically as the ideal theory requires. Such errors can alter the experimental results and – crucially – make the experiment easier to simulate classically, which would potentially negate the intended quantum advantage. The researchers’ analysis helps to distinguish between imperfections that do not fundamentally negate the complexity of the sampling problem and those that are potentially more harmful.
This distinction is valuable as it can guide the development of future photonic quantum technologies. Rather than treating every imperfection as equally serious, experimental teams can focus their resources on reducing the effects most likely to undermine computational complexity. In this sense, the study provides both the assurance that even imperfect devices can still be significant, and a practical roadmap for improving the next generation of quantum-optical sampling experiments.
The research was funded by the Federal Ministry of Research, Technology and Space (BMFTR) as part of the PhoQuant project (grant number 13N16103).
The article has been published in the journal *Physical Review A*: https://doi.org/10.1103/pddb-5z5q
A freely available version can be downloaded from the ArXiv preprint server: doi.org/10.1103/pddb-5z5q
