Skip to main content

M09 · Information Theory

A free, illustrated L2 to L3 course of 8 modules in the Mathematics track. Modules include: Surprise and entropy; Joint entropy, conditional entropy and mutual information; KL divergence and cross-entropy.

M09 L2 to L3 · Intermediate to Advanced 8 modules · about 32 lessons Planned · Ring 3

Part of the Mathematics track. Ships in Ring 3 (Breadth) of the roadmap.

Modules

  1. Surprise and entropy
  2. Joint entropy, conditional entropy and mutual information
  3. KL divergence and cross-entropy
  4. Source coding and compression
  5. Channel capacity
  6. Maximum entropy and exponential families
  7. Minimum description length and model selection
  8. Information theory inside ML: the ELBO and contrastive bounds

Signature illustrations

The pictures a learner should still see in their head a year later:

Entropy as average number of yes-or-no questions, KL asymmetry plots, information Venn diagrams, Huffman trees being built.

Status

This course is planned, not written yet. It ships with Ring 3 (Breadth).

Every lesson will follow the same twelve-part blueprint with at least three purpose-made figures. Watch the repository to hear when modules ship, or help write and review it.

Back to top