From Quantum Supremacy to Quantum Utility: Reservoir Computing with Boson Sampling
Boson sampling has been central to demonstrations of quantum supremacy, yet its lack of practical applications has limited its broader impact. Here we show how boson sampling can be transformed from a benchmark of computational hardness into a versatile resource for machine learning. By harnessing the complex, high-dimensional dynamics of random interferometers, we realize quantum reservoir computing capable of tackling image recognition tasks. Our results demonstrate that even modest-scale boson sampling systems can achieve non-trivial performance, opening a pathway from supremacy-driven demonstrations to genuine quantum utility.
[A. Sakurai, et.al , Quantum optical reservoir computing powered by boson sampling, Optica Quantum 3(3), 238-245 (2025)]