Specific Emitter Identification with Limited Labelled Signals Based on Variational Autoencoder Embedded in Information-Maximising Generative Adversarial Network and Gradient Penalty

Table 1

Network structure and hyperparameters.

Encoder E
Conv 2D	, , Tanh, batch norm
Conv 2D	, , Tanh, batch norm
Dense	, ReLU, batch norm
Generator G
Dense	, LeakyReLU, batch norm
Reshape
Conv 2D	, , Tanh, batch norm
Conv 2D	, , Tanh, batch norm
Conv 2D	, , Tanh, batch norm
Discriminator D & auxiliary classifier Q
Conv 2D	, , ReLU, batch norm, Maxpooling
Conv 2D	, , ReLU, batch norm, Maxpooling
Conv 2D	, , ReLU, batch norm
Discriminator D		Auxiliary classifier Q
Dense	, ReLU	Dense	, ReLU
Dense	, sigmoid	Dense	, ReLU
		Dense	, softmax
Hyperparameters		Value
Optimiser		Adam
Learning rate		0.0001
Epochs		400
Minibatch		64
Dropout rate		0.5