Adversarial Learning with Ladder Networks

Abstract

[EN] In this MT we are going to asses the inclusion of adversarial generative models into the Ladder semi-supervised networks. The implementation and experiments are carried out with Theano. Therefore the adversarial generative model has to be included in the derivation graph.

[ES] En este TFM se va a analizar el impacto de introducir modlos generativos de muestras adversarias en las redes profundas semi-supervisadas del tipo "ladder". La implementación y experimentos se realizará en Theano con lo que se tendrá que incluir el modelo generativo adversario en el grado de derivación.