Mastering Machine Learning for Penetration Testing

The discriminator

The discriminator is simply a classifier trained with supervised learning techniques to check if the image is real (1) or fake (0). It is trained by both the MNIST dataset and the generator samples. The discriminator will classify the MNIST data as real, and the generator samples as fake:

discriminator = Sequential([Dense(128, input_shape=(784,)),LeakyReLU(alpha=0.01),Dense(1),Activation('sigmoid')], name='discriminator')

By connecting the two networks, the generator and the discriminator, we produce a generative adversarial network:

gan = Sequential([generator,discriminator])

This is a high-level representation of a generative adversarial network:

To train the GAN, we need to train the generator (the discriminator is set ...