Effect of Input Layer Signal Polarization on the Dynamics of Optical Neural Networks

Abstract

The polarization encoding-based optical validation and security verification approach is provided in this paper. This technique involves simulating information optically and bonding it to a polarization-encoded mask, such as a biological order or a reaction. The linear polarizers that make up the polarization-encoded mask are positioned at random. The polarization-encoded signal is the name given to this composite signal. In this simulation study, a primary optical neural network adapting a light brain technology is proposed theoretically based on a feed-forward model. Calibration of the nonlinear behavior in such a network is assumed by a semiconductor laser of the Distributed Feedback (DFB) type. Four laser networks are constructed as three influencers, followed by one embedding laser followers. Each of the influencer’s lasers has a different wavelength frequency and polarization (30-60–90) degree, respectively, and then combines the signal with WDM for the last laser. With each value, from the last values, of the polarization effect after this effect, the results indicated that these values would present the greatest weight of spikes and chaotic behavior for the uploaded virtual message.