AI Glossary Term
What does “Continual Learning” mean?
Training approaches that let a model keep learning from new data without forgetting prior knowledge.
How Continual Learning works
Example
A model trained to recognize cats is then fine-tuned only on birds. Without precautions it may become worse at cats, a problem called catastrophic forgetting. Continual learning methods, such as replaying a sample of old data, try to avoid this.
Why it matters
Real-world data keeps changing, and retraining from scratch is expensive. Systems that can keep updating without forgetting would be cheaper and more current.
Common questions
What is catastrophic forgetting?
When learning new data overwrites weights that stored earlier knowledge, so performance on old tasks drops sharply.
How is it different from fine-tuning?
Fine-tuning adapts to a new task and may forget old ones. Continual learning specifically aims to keep old abilities while adding new ones.
Do large language models learn continually?
Typically no. They are trained in stages and then released as fixed versions, with updates arriving as new versions.