Machine Learning Foundations

Learning Rate, explained

The learning rate controls how large each optimization update is during training.
Back to Glossary

Part of the Semantic Notion technical glossary.

Short Definition

The learning rate controls how large each optimization update is during training.

Intuition

A small learning rate learns slowly but steadily. A large learning rate moves faster but can bounce around or diverge.

Technical Definition

The learning rate is the scalar step size applied to gradient-based parameter updates.

Example

If training loss explodes, lowering the learning rate is often one of the first checks.

Common Misunderstandings

There is no universal best learning rate.
Learning-rate schedules can matter as much as the initial value.

Start here

Need the broader concept map?

Return to the glossary index to move from this definition into adjacent AI, ML, mathematics, and computer science terms.