Entropy lies between zero and the logarithm of the order
Lax263307.EntropyBounds · concepts/Lax263307/EntropyBounds.lean · lax-263307
No public endorsements yet.
Loading review…
Sign in with ORCIDNatural Language Statement
Theorem
For every probability vector on vertices,
Concept map
Lean source view on GitHub
| 1 | import Lax263307.Entropy |
| 2 | |
| 3 | /-! |
| 4 | --- |
| 5 | title: Entropy lies between zero and the logarithm of the order |
| 6 | type: theorem |
| 7 | --- |
| 8 | For every probability vector on vertices, |
| 9 | |
| 10 | |
| 11 | # Attribution |
| 12 | |
| 13 | Bonnet–Chang, proof of Lemma 12. The upper bound follows from concavity |
| 14 | of , with equality at the uniform distribution. |
| 15 | -/ |
| 16 | |
| 17 | namespace Lax263307.EntropyBounds |
| 18 | |
| 19 | open Entropy |
| 20 | |
| 21 | /-- The entropy bounds used to limit successive conditioning steps. -/ |
| 22 | axiom entropy_bounds {n : ℕ} (π : Fin n → ℝ) (probability : IsProbability π) : |
| 23 | 0 ≤ entropy π ∧ entropy π ≤ Real.logb 2 n |
| 24 | |
| 25 | end Lax263307.EntropyBounds |
| 26 |
Attribution
Bonnet–Chang, proof of Lemma 12. The upper bound follows from concavity of , with equality at the uniform distribution.
Discussion
Ask a question or add context. Endorsements and structured flags are kept in the review panel above.
0 comments