AFTD Auto-Formalizing Theoretical Domains

dephased_path_entropy

def

For n : ℕ, q : Fin n → ℝ, and S : Finset (Fin n), dephased_path_entropy n q S is defined as (S.card : ℝ) - (free.card : ℝ) plus the sum over all free connected components c of dephased_path_aux_graph n S of binaryEntropy ((1 - ∏ u ∈ Finset.univ.filter (fun u => G.connectedComponentMk u = c), q u.1) / 2).

Statement

noncomputable def dephased_path_entropy (n : ℕ) (q : Fin n → ℝ) (S : Finset (Fin n)) : ℝ
Defined in
arXiv:2511.19585, Monogamy of Mutual Information in Graph States
Verified
02 Nov 2026
Axioms
none listed
Built on 3
  • binaryEntropyThe binary entropy function \[ H(x) \;=\; -x\log_2 x - (1-x)\log_2(1-x), \] using the base-$2$ logarithm, with the convention…
  • dephased_path_aux_graphFor n : ℕ and a finset S : Finset (Fin n), the auxiliary graph dephased_path_aux_graph n S is the SimpleGraph on the subtype S in which…
  • dephased_path_free_componentFor n : ℕ and S : Finset (Fin n), a connected component c of dephased_path_aux_graph n S is free if every vertex u : S in the component…
Used by 1
  • dephased_path_tripartite_infoFor n : ℕ, q : Fin n → ℝ, and finsets A, B, C : Finset (Fin n), the tripartite information dephased_path_tripartite_info n q A B C is…

Read back from the Lean

dephased_path_entropy n q S is a noncomputable definition that assigns a real number to an integer n : ℕ, a weighting function q : Fin n → ℝ, and a subset of vertices S : Finset (Fin n). It computes: $$|S| - |\mathrm{free}| + \sum_{c \in \mathrm{free}} \operatorname{binaryEntropy}\left(\frac{1 - \prod_{u \in c} q(u)}{2}\right)$$ where: 1. $G$ is the auxiliary simple graph on the vertex set $S$ in which two distinct vertices $u, w \in S$ are adjacent if and only if they share a common neighbor $v \in \mathrm{Fin}\ n \setminus S$ in the path graph $P_n$. 2. A vertex $u \in S$ is *marked* if there exists some $v \in \mathrm{Fin}\ n \setminus S$ whose only neighbor in $S$ (in $P_n$) is $u$. 3. A connected component $c$ of $G$ is *free* if no vertex in $c$ is marked. The finset of all such components is denoted $\mathrm{free}$. 4. For each component $c \in \mathrm{free}$, the product $\prod_{u \in c} q(u)$ is taken over all vertices $u \in S$ belonging to $c$. 5. $\operatorname{binaryEntropy}(x) = -x \log_2(x) - (1-x) \log_2(1-x)$.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

Lean source view module on GitHub

/-- Entropy set function for a dephased path graph state under independent Z-flips. -/
noncomputable def dephased_path_entropy (n : ℕ) (q : Fin n → ℝ) (S : Finset (Fin n)) : ℝ := open scoped Classical in
  let G := dephased_path_aux_graph n S
  let free : Finset G.ConnectedComponent := Finset.univ.filter (fun c => dephased_path_free_component n S c)
  (S.card : ℝ) - (free.card : ℝ) +
    ∑ c ∈ free, binaryEntropy ((1 - ∏ u ∈ Finset.univ.filter (fun (u : S) => G.connectedComponentMk u = c), q u.1) / 2)
See it in the graphDiscuss this resultChallenge it