← Back

note / modern mathematics

Session 10 — Shannon Information

Shannon Information

Central Question

How does uncertainty become a statement about communicating and compressing messages?

Why This Matters

Information theory connects a measure of uncertainty with operational limits on description length and communication, giving entropy a concrete interpretation.

H(X)=xp(x)logp(x)H(X)=-\sum_x p(x)\log p(x)

Learning Prompt

Develop Shannon entropy H(X)=xp(x)logp(x)H(X)=-\sum_x p(x)\log p(x) as the expected information of a discrete random variable, building from surprising outcomes and repeated messages rather than announcing a formula. Work through a biased binary source and motivate the connection with lossless coding and typical sequences, stating the assumptions and meaning of an asymptotic limit. Introduce conditional entropy and mutual information through a concrete dependent pair. Make me distinguish entropy from semantic meaning and from the length of one particular message. Use conceptual predictions to connect uncertainty, compression and dependence.

Ideas

Examples

Questions

Connections

One Thing That Surprised Me

Further Rabbit Holes