1 result for “representation-learning”
An autoencoder is a type of artificial neural network trained to reconstruct its input through a compressed internal representation, used for dimensionality reduction, feature learning, and anomaly detection.