Autoencoder

Models and Architectures

A neural network that learns to compress data down to its essence and reconstruct it - useful for spotting anything unusual.

An autoencoder squeezes its input through a narrow bottleneck and then tries to rebuild the original from that compressed summary. To succeed it must learn what really matters in the data and discard the noise.The practical payoff: anything the network reconstructs badly is unusual - which makes autoencoders a favorite tool for anomaly detection in fraud, manufacturing, and network security, alongside uses in denoising and compression.