A photonic complex perceptron for ultrafast data processing

Abstract

In photonic neural network a key building block is the perceptron. Here, we describe and demonstrate a complex-valued photonic perceptron that combines time and space multiplexing in a fully passive silicon photonics integrated circuit. An input time dependent bit sequence is broadcasted into a few delay lines where the relative phases are trained by particle swarm algorithms toward the given task. Since only the phases of the propagating optical modes are trained, signal attenuation in the perceptron due to amplitude modulation is avoided. The perceptron performs binary pattern recognition and few bit delayed XOR operations up to 16 Gbps (limited by the used electronics) with Bit Error Rates as low as $10^{-6}$. The perceptron is fully integrated, silicon based, scalable, and can be used as a building block in large neural networks.

13 pages, 9 figures

References in corpus (4)

• Photonics for artificial intelligence and neuromorphic computing
• 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
• Do Reservoir Computers Work Best at the Edge of Chaos?
• Local homeostatic regulation of the spectral radius of echo-state networks

Cited by in corpus (4)

• Photonic neural networks based on integrated silicon microresonators
• Equalization of a 10 Gbps IMDD signal by a small silicon photonics time delayed neural network
• A Silicon Photonic Neural Network for Chromatic Dispersion Compensation in 20 Gbps PAM4 Signal at 125 km and Its Scalability up to 100 Gbps
• Nonlinear Distortion Equalization in Multi-Span Optical Links Via a Feed-Forward Photonic Neural Network