Overview
An autoencoder consists of an Encoder that compresses the input into a lower-dimensional 'latent space' and a Decoder that tries to reconstruct the original input from that compressed representation.
Purpose
- Dimensionality Reduction: Similar to PCA but non-linear.
- Denoising: Learning to remove noise from images or audio.
- Feature Learning: Discovering the most important characteristics of a dataset.
Variants
Variational Autoencoders (VAEs) are a popular generative version used to create new data (like faces or music).