Hoi-Fung Chau

Photons play an important role in quantum information processing as they are easy to manipulate locally and transfer over a long distance. Photonic qubits are widely used in tasks such as linear optical quantum computing (LOQC), quantum sensing/metrology, and quantum communication. However, to date, efficient high-speed single photon sources are still difficult to make. Here, we propose that we can use "classical" phase-randomized coherent states, combined with post-processing, to perform various quantum information processing tasks. Specifically, we divide the tasks into two scenarios: ones with a circuit of a known Hilbert space dimension, describable by a unitary matrix, such as LOQC and metrology, as well as ones with an unknown channel, such as quantum communication. We propose methods that can hugely improve the numerical precision and applicable dimensions in both scenarios, including a machine learning method for the former and a linear interpolation method for the latter, opening up a wide variety of applications that can be implemented with easily attainable coherent light sources and threshold detectors, such as quantum metrology or universal fully-passive state preparation.