Phase signal embedment in densely frequency-multiplexed coherent neural networks
Amornrat Limmanee, Sotaro Kawata, Akira Hirose · Adaptive Optics: Analysis and Methods/Computational Optical Sensing and Imaging/Information Photonics/Signal Recovery and Synthesis Topical Meetings on CD-ROM · 2005
We propose a coherent neural network whose synaptic weights are embedded densely in volume hologram using a semiconductor laser so that the behavior is controlled by direct lightwave carrier-frequency modulation. A frequency-dependent learning obeys the complex-valued Hebbian rule. Experiments demonstrate that the system learns different output phase values dependently on the carrier frequency.