Autoencoder

An autoencoder is a type of <a href="/facts/Artificial_neural_network/6V1jMlkx">artificial neural network</a> used to learn <a href="/facts/Feature_learning/6GYV0mux">efficient codings</a> of unlabeled data (<a href="/facts/Unsupervised_learning/8a6f61MC">unsupervised learning</a>). An autoencoder learns two functions: an encoding function that transforms the input data, and a decoding function that recreates the input data from the encoded representation. The autoencoder learns an efficient representation (encoding) for a set of data, typically for <a href="/facts/Dimensionality_reduction/XoRpcrB0">dimensionality reduction</a>, to generate lower-dimensional embeddings for subsequent use by other <a href="/facts/Machine_learning/e0w0XJTu">machine learning</a> algorithms.
Variants exist which aim to make the learned representations assume useful properties. Examples are regularized autoencoders (sparse, denoising and contractive autoencoders), which are effective in learning representations for subsequent <a href="/facts/Statistical_classification/jXXHRkXR">classification</a> tasks, and <a href="/facts/Variational_autoencoder/I5wBbYpR">variational autoencoders</a>, which can be used as <a href="/facts/Generative_model/JUgExNIP">generative models</a>. Autoencoders are applied to many problems, including <a href="/facts/Facial_recognition_system/mtVxaDln">facial recognition</a>, <a href="/facts/Feature_(computer_vision)/GdfWRp5U">feature detection</a>, <a href="/facts/Anomaly_detection/wjEGMGGz">anomaly detection</a>, and <a href="/facts/Word_embedding/7uRcBPqo">learning the meaning of words</a>. In terms of <a href="/facts/Synthetic_data/MVSNrFPI">data synthesis</a>, autoencoders can also be used to randomly generate new data that is similar to the input (training) data.

Autoencoder open-in-new

Autoencoder