Add to BibSonomyAbstract
We experimentally demonstrate the first all-optical Recurrent-Neural-Network (RNN) with a gating circuit and a sigmoid activation unit. By introducing a gating circuit to the RNN we implemented the “forget” function, resembling an all-optical Gated-Recurrent-Unit (GRU).
© 2020 The Author(s)
PDF ArticleMore Like This
All-optical recurrent neural network with sigmoid activation function
George Mourgias-Alexandris, George Dabos, Nikolaos Passalis, Anastasios Tefas, Angelina Totovic, and Nikos Pleros
W3A.5 Optical Fiber Communication Conference (OFC) 2020
Supercontinuum spectral evolution prediction by recurrent neural network
Lauri Salmela, Coraline Lapre, John M. Dudley, and Goery Genty
STu4H.6 CLEO: Science and Innovations (CLEO:S&I) 2020
Recurrent Neural Networks for Short-Term Forecast of Lightpath Performance
Sandra Aladin, Stéphanie Allogba, Anh Vu Stephan Tran, and Christine Tremblay
W2A.24 Optical Fiber Communication Conference (OFC) 2020