Memory in Machine Learning

Machine learning uses memory vocabulary for learned parameters, sequence state, replay buffers, external memory modules, retrieval systems, and the challenge of learning new information without erasing old information.

Core terms: learning, catastrophic forgetting, continual learning, experience replay, recurrent neural network, long short-term memory, Hopfield network, Neural Turing Machine, Differentiable Neural Computer.

Terms that change meaning across fields: memory, retention, forgetting, state, retrieval, adaptation.

Return to the Canonical Memory Glossary.