Glossary
What is Information theory? Definition and meaning
Entropy, mutual information and KL divergence, which underlie many training losses.
Definition
Information theory: Entropy, mutual information and KL divergence, which underlie many training losses.
In short: Measuring information, entropy and divergence.
When it is used
Loss design, contrastive objectives and compression arguments.
Learn more
The full guide to Information theory explains how it works, when to use it and how it compares with the alternatives, with examples. The glossary lists every term on the site from A to Z.
Previous term: Informatica EDC · Next term: Infrastructure as code