« Back to Glossary Index

A technique used to train Artificial Neural Networks by computing the gradient required to adjust the network’s weights.