REVISION AND PRACTICE
Check your understanding
- What is the difference between n and d in a dataset?
- Why does “identically distributed” not mean identical samples?
- What distribution does have, and what determines it?
- Why is argmin not the same object as a minimum value?
- What is the correct direction of the convexity inequality?
- Which extra condition on f makes Jensen imply nonnegative f-divergence?
- Why can KL divergence fail to be a distance metric?
About the examples: Numerical exercises and prerequisite detours are original teaching aids. They clarify chapter material rather than introducing new syllabus topics.