Fig. 2: Demonstration of hybrid and classical networks learning to classify the synthetically generated dataset. | npj Unconventional Computing

Fig. 2: Demonstration of hybrid and classical networks learning to classify the synthetically generated dataset.

From: Hybrid quantum-classical photonic neural networks

Fig. 2

a The maximum output activations for an untrained 120-parameter hybrid network, network A. The x-axis represents the predicted class, while the color indicates the true class. The inset shows the maximum output activations for all samples predicted as class three. b The maximum output activations for an untrained 124-parameter classical network, B. c The maximum output activations for network A after training. d The maximum output activations for a trained network B. e The training and validation accuracy curves for networks A and B.

Back to article page