Indistinguishable Photon Source in the 1550-nm Band Optimized by Machine Learning

We report a recipe for optimized poling design to generate spectrally-pure photons around 1550 nm. Our approach hinges on machine-learning techniques without requiring group-velocity matching. © 2019 The Author(s)

Indistinguishable photons are key elements for Boson Sampling to demonstrate quantum supremacy [1]–[3]. They are also essential ingredients in entanglement swapping to enable large-scale quantum networks. The University of Arizona (UA) is constructing a quantum-network testbed that utilizes entangled photons to connect different buildings and diverse disciplines, as illustrated in Fig. 1. To this end, we are building robust indistinguishable photon sources operating in the 1550-nm band to readily access the abundant telecommunication resources.