A photonic CNN performs MNIST classification entirely in the optical domain at 94% accuracy using MZI meshes and microring nonlinearities, with hybrid ex-situ digital twin training followed by in-situ SPSA fine-tuning.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.ET 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
contest 1representative citing papers
citing papers explorer
-
Photonic convolutional neural network with pre-trained in-situ training
A photonic CNN performs MNIST classification entirely in the optical domain at 94% accuracy using MZI meshes and microring nonlinearities, with hybrid ex-situ digital twin training followed by in-situ SPSA fine-tuning.