Photonic Max-Pooling for Deep Neural Networks Using a Programmable Photonic Platform
Farshid Ashtiani, Mehmet Berkay On, David Sánchez-Jácome, Daniel Pérez, S. J. Ben Yoo, Andrea Blanco‐Redondo · 2023
We propose a photonic max-pooling architecture for photonic neural networks which is compatible with integrated photonic platforms. As a proof of concept, we have experimentally demonstrated the max-pooling function on a programmable photonic platform consisting of a hexagonal mesh of Mach-Zehnder interferometers.