Experimental Quantum Generative Adversarial Networks for Image Generation

Unknown author(s)

Physical Review Applied · 2021 · 255 citations · 41 references

DOIFull text

Open access

Concepts

Abstract

Quantum machine learning is expected to be among the first practical applications of near-term quantum devices. Whether quantum generative adversarial networks (quantum GANs) implemented on near-term devices can actually solve real-world learning tasks, however, has remained unclear. The authors narrow this knowledge gap by designing a flexible quantum GAN scheme, and realizing this scheme on a superconducting quantum processor. Their system learns and generates images of real-world handwritten numerals, and exhibits competitive performance with classical GANs. This work opens up an avenue for exploring quantum advantage in various machine-learning tasks.

References

41