.png){kind=icon}
.png){kind=icon}
Return
Quantum-Enabled Nonlocal Optical Convolutional Processing Using Energy-Time Entanglement
.png){kind=icon}
PRE
AI
DOI：10.1002/lpor.202502133.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Conventional optical convolution processors face fundamental limitations in handling spatially separated signals and operating at ultra-low-light levels down to the single-photon regime. To overcome these constraints, we introduce a quantum-enabled nonlocal optical convolutional processor utilizing energy-time entangled photon pairs as the optical carrier. As a proof-of-principle demonstration, we achieve image recognition of a handwritten digit 0 from MNIST dataset. In our system, input data and reconfigurable kernels (2x2/3x3) are encoded onto the idler photons, while a dispersion element acts on the signal photons. Leveraging quantum temporal correlation between the photon pairs, the encoded data is nonlocally mapped onto the signal photons for a dispersion-induced convolution. To achieve precise spectral tailoring of the optical carrier, we employ multi-window post-selection on the biphoton coincidence distribution. This technique allows the convolution result to be read out from the heralded signal photons, bypassing the local encoding required in conventional processors. Without the local encoding required in conventional processors. This approach performs information routing entirely within the optical domain, avoiding inefficient optical-electrical-optical conversion and showing significant potential in quantum energy-time entanglement, nonlocality, and optical convolutional processor communication networks.
Journal
L
IF:
10
Papers:
1.1K
Citations:
1