Autoencoder 07/06/2024 Autoencoder Keywords Unsupervised learning Dimensionality reduction = Representation learning = Feature extraction Generative model learning Properties Input and output are the same size It consists of two parts: Encoder and a Decoder autoencoder Hidden Bottleneck Layer architecture The term for Code in the above figure is same as z, latent variable, feature ... Encoder performs a role as dimensionality reduction Decoder performs a role as a generative model It reduces dimensions and then reconstructs them, learning the core features of the data and generating a compressed representation Simple implementation