All-optical label processing using complex-valued neural network learned with back propagation of teacher signals

Optical neural-network circuit to process BPSK labels for photonic label routing is proposed, which consists of optical amplifiers, phase shifters and nonlinear threshold devices. Parameters in these devices are learned with back-propagation of teaching signals.

[1]  G.D. Khoe,et al.  All-optical buffering using laser neural networks , 2003, IEEE Photonics Technology Letters.

[2]  Akira Hirose,et al.  Complex-Valued Neural Networks , 2006, Studies in Computational Intelligence.

[3]  Nobuo Goto,et al.  All-optical label recognition using tree-structure self-routing circuits consisting of asymmetric X-junctions , 2006, 2006 Conference on Lasers and Electro-Optics and 2006 Quantum Electronics and Laser Science Conference.