Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 8.1.1 — Lieb concavity

Proved
TroppMatrixConcentration.lieb_concavity

by tc · Oct 7, 2026 · Mathlib 0df444a (Lean v4.33.1)

matrix-concentrationrandom-matrices

Fix a Hermitian complex d×dd\times dd×d matrix HHH, with d≥1d\ge1d≥1. The real-valued function

A⟼tr⁡exp⁡(H+log⁡A)A\longmapsto\operatorname{tr}\exp(H+\log A)A⟼trexp(H+logA)

is concave on the convex cone of positive-definite complex d×dd\times dd×d matrices, viewed as a real convex set. The logarithm is the spectral matrix logarithm. This concavity result is a reusable matrix-analysis input to the concentration theory.

Preamble
import Definitions.Def_TroppMatrixConcentration_probability
import Definitions.Def_TroppMatrixConcentration_dilation
import Mathlib.Analysis.Convex.Function

open MeasureTheory ProbabilityTheory
open scoped Matrix.Norms.L2Operator ComplexOrder

Formal statement
namespace TroppMatrixConcentration

theorem lieb_concavity {d : ℕ} [NeZero d]
    (H : Matrix (Fin d) (Fin d) ℂ) (hH : H.IsHermitian) :
    ConcaveOn ℝ {A : Matrix (Fin d) (Fin d) ℂ | A.PosDef}
      (fun A => traceExp (H + matrixLog A)) := by sorry

end TroppMatrixConcentration
Source
Joel A. Tropp, An Introduction to Matrix Concentration Inequalities, arXiv:1501.01571v1 (7 January 2015); https://arxiv.org/abs/1501.01571v1; Theorems 3.4.1 and 8.1.1, printed pp. 35 and 119; proof development Chapter 8.
Read-back

What the Lean code literally says, in plain math · gpt-6

For every natural number d≠0d\ne 0d=0 and every complex d×dd\times dd×d matrix HHH satisfying H∗=HH^*=HH∗=H, let Pd\mathcal P_dPd​ be the set of Hermitian matrices AAA such that x∗Axx^*Axx∗Ax is a strictly positive real number for every nonzero vector x∈Cdx\in\mathbb C^dx∈Cd. Define log⁡A\log AlogA by applying the real logarithm to AAA through the continuous functional calculus; for A∈PdA\in\mathcal P_dA∈Pd​, this means log⁡A=Udiag⁡(log⁡λ1,…,log⁡λd)U∗\log A=U\operatorname{diag}(\log\lambda_1,\ldots,\log\lambda_d)U^*logA=Udiag(logλ1​,…,logλd​)U∗ whenever A=Udiag⁡(λ1,…,λd)U∗A=U\operatorname{diag}(\lambda_1,\ldots,\lambda_d)U^*A=Udiag(λ1​,…,λd​)U∗ is a unitary diagonalization, with all λi>0\lambda_i>0λi​>0. Define exp⁡M=∑k=0∞Mk/k!\exp M=\sum_{k=0}^{\infty}M^k/k!expM=∑k=0∞​Mk/k! and FH(A)=Re⁡tr⁡(exp⁡(H+log⁡A))F_H(A)=\operatorname{Re}\operatorname{tr}(\exp(H+\log A))FH​(A)=Retr(exp(H+logA)), where the trace is the sum of the diagonal entries. Then Pd\mathcal P_dPd​ is convex under real scalar combinations, and for every A,B∈PdA,B\in\mathcal P_dA,B∈Pd​ and every pair of real numbers a,b≥0a,b\ge 0a,b≥0 satisfying a+b=1a+b=1a+b=1, one has aA+bB∈PdaA+bB\in\mathcal P_daA+bB∈Pd​ and aFH(A)+bFH(B)≤FH(aA+bB)aF_H(A)+bF_H(B)\le F_H(aA+bB)aFH​(A)+bFH​(B)≤FH​(aA+bB). The quantifiers include d=1d=1d=1, A=BA=BA=B, and the endpoint weights (a,b)=(0,1)(a,b)=(0,1)(a,b)=(0,1) and (1,0)(1,0)(1,0); dimension zero and matrices that are singular or merely positive semidefinite are excluded from the asserted domain, and no concavity assertion is made outside Pd\mathcal P_dPd​.

Human review
  • Endorsed by Shuze Chen · Oct 7, 2026

    Confirmed by the moderator at approval.

  • Endorsed by tc · Oct 7, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me