Title: Boson Sampling in Low-depth Optical Systems

Abstract: Optical losses are the main obstacle to demonstrating a quantum advantage via boson sampling without leaving open the possibility of classical spoofing. We propose a method for generating low-depth optical circuits suitable for boson sampling with very high efficiencies. Our circuits require only a constant number of optical components (namely three) to implement an optical transformation suitable for demonstrating a quantum advantage. Consequently, our proposal has a constant optical loss regardless of the number of optical modes. We argue that sampling from our family of circuits is computationally hard by providing numerical evidence that our family of circuits converges to that of the original boson sampling proposal in the limit of large optical systems. Our work opens a new route to demonstrate an optical quantum advantage.