Machine unlearning – Definition & Meaning

Definition of machine unlearning

: the process of removing the influence of specified training examples, knowledge, or capabilities from a trained machine-learning model, ideally without retraining the model from the beginning
- The company used machine unlearning after a customer requested that her records no longer influence the recommendation model.
- Researchers tested whether the unlearned model still reproduced facts from the removal set.

Origin & history

first known use2015

The modern term was introduced in the 2015 IEEE Symposium on Security and Privacy paper Towards Making Systems Forget with Machine Unlearning. It first focused on efficiently removing data from learned systems and later expanded to foundation models, where researchers also study removing concepts, copyrighted material, unsafe capabilities, or personal information.

Test yourself

Which of these is the meaning of machine unlearning?

Related terms

training data the collection of examples (text, images, code, audio, and so on) that a model learns from fine-tuning further training of an existing model on a smaller, focused dataset so it gets better at a… data poisoning an attack that inserts, changes, or selectively removes examples in a machine-learning… model collapse a gradual loss of quality and diversity that happens when models are trained on content… alignment the work of making an AI system's goals and behavior match what its designers and users… large language model a very large neural network trained on huge amounts of text to predict the next token, which…

Cite this entry

"Machine unlearning." AI Dictionary, Dadgogo, https://dadgogo.com/dictionary/machine-unlearning/. Accessed 8 Oct. 2026.