Autoencoder

An Autoencoder is a specialized type of artificial neural network primarily used for unsupervised learning, designed to learn efficient codings or representations of unlabeled data. Its architecture consists of two main parts: an "encoder" that compresses the input data into a lower-dimensional representation (often called the latent space or bottleneck layer), and a "decoder" that attempts to reconstruct the original input from this compressed representation. The network is trained to minimize the "reconstruction error," the difference between the original input and its reconstructed output.