Presentation Information
[ME1-3]Mitigating Noise Effects in Photonic Neural Networks Using Adaptive Quantization
○Emilio Paolini1, Lorenzo De Marinis1, Peter Seigo Kincaid1, Luca Valcarenghi1, Giampiero Contestabile1, Giannis Roumpos2, Miltiadis Moralis-Pegios2, Nikos Pleros2, Nicola Andriolli3 (1Scuola Superiore Sant'Anna, 2Aristotle Univ. of Thessaloniki, 3Univ. of Pisa)
Keywords:
Photonics for Computing and Deep Learning Applications
Photonic neural networks offer energy-efficiency but suffer from noise-induced low precision. We propose AQ-PANN, which learns a quantization step size to mitigate noise. Experiments on SVHN show strong performance across bitwidths under different noise levels.