Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Block-diagonal Schatten embedding and contractive pinching

Proved
QAlgorithms.Papers.Independent.Support21.P15.schatten_blockDiagonal

by Goku · Oct 21, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let 0<p<∞0<p<\infty0<p<∞ and let n,m≥1n,m\ge1n,m≥1 be integers. Write SpmS_p^mSpm​ for the complex m×mm\times mm×m matrices with Schatten quasi-norm ∥A∥p=(Tr⁡((A∗A)p/2))1/p\|A\|_p=(\operatorname{Tr}((A^*A)^{p/2}))^{1/p}∥A∥p​=(Tr((A∗A)p/2))1/p. The complex-linear map B(A1,…,An)=diag⁡(A1,…,An)B(A_1,\ldots,A_n)=\operatorname{diag}(A_1,\ldots,A_n)B(A1​,…,An​)=diag(A1​,…,An​) is injective and satisfies

∥B(A1,…,An)∥p=(∑i=1n∥Ai∥pp)1/p.\|B(A_1,\ldots,A_n)\|_p=\left(\sum_{i=1}^n\|A_i\|_p^p\right)^{1/p}.∥B(A1​,…,An​)∥p​=(i=1∑n​∥Ai​∥pp​)1/p.

For p≥1p\ge1p≥1, deleting the off-diagonal blocks defines a complex-linear projection PPP onto the range of this same inclusion, with ∥P(A)∥p≤∥A∥p\|P(A)\|_p\le\|A\|_p∥P(A)∥p​≤∥A∥p​ for every matrix AAA. Thus this range is 1-complemented in SpnmS_p^{nm}Spnm​.

This identifies finite p-sums of Schatten blocks with concrete block-diagonal matrix subspaces. Formalization Note The target matrix rows and columns are indexed by Fin(n)×Fin(m)\mathrm{Fin}(n)\times\mathrm{Fin}(m)Fin(n)×Fin(m); the Schatten formula is reused from the pinned accepted library.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support21_P15
Formal statement
noncomputable section

open scoped BigOperators

open Filter

namespace QAlgorithms.Papers.Independent.Support21.P15

/-- Unnumbered block-diagonal observation, arXiv:math/0307169v1, PDF p.14:
ell_p^n(S_p^m) embeds isometrically into S_p^(nm), with a contractive projection
onto that same range when p≥1. The Schatten size is the accepted all-matrix
trace-of-power formula, not the ambient entrywise matrix norm. -/
theorem schatten_blockDiagonal (p : ℝ) (hp : 0 < p)
    (n m : ℕ) (hn : 1 ≤ n) (hm : 1 ≤ m) :
    Function.Injective (schattenBlockInclusion n m) ∧
    (∀ x : Fin n → Matrix (Fin m) (Fin m) ℂ,
      QAlgorithms.Papers.Independent.Information100.P50.schattenNorm (ENNReal.ofReal p)
        (schattenBlockInclusion n m x) =
      finiteLpSize p
        (QAlgorithms.Papers.Independent.Information100.P50.schattenNorm (ENNReal.ofReal p)) x) ∧
    (1 ≤ p →
      (schattenBlockProjection n m).comp (schattenBlockProjection n m) =
        schattenBlockProjection n m ∧
      LinearMap.range (schattenBlockProjection n m) =
        LinearMap.range (schattenBlockInclusion n m) ∧
      ∀ x : Matrix (Fin n × Fin m) (Fin n × Fin m) ℂ,
        QAlgorithms.Papers.Independent.Information100.P50.schattenNorm (ENNReal.ofReal p)
          (schattenBlockProjection n m x) ≤
        QAlgorithms.Papers.Independent.Information100.P50.schattenNorm (ENNReal.ofReal p) x) := by sorry

end QAlgorithms.Papers.Independent.Support21.P15

end
Source
On subspaces of non-commutative L_p-spaces, arXiv:math/0307169v1, PDF p. 14, Unnumbered block-diagonal Schatten embedding
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

For every real p>0p>0p>0 and integers n,m≥1n,m\geq1n,m≥1, define, for each finite square complex matrix AAA,

Sp(A)=(Re⁡tr⁡ ⁣[(A∗A)p/2])1/p,S_p(A)= \left(\operatorname{Re}\operatorname{tr}\!\left[(A^*A)^{p/2}\right]\right)^{1/p},Sp​(A)=(Retr[(A∗A)p/2])1/p,

where A∗A^*A∗ is the conjugate transpose and the matrix power is defined by continuous functional calculus using the real function t↦tp/2t\mapsto t^{p/2}t↦tp/2. Let JJJ send a family x=(xa)a<nx=(x_a)_{a<n}x=(xa​)a<n​ of m×mm\times mm×m matrices to the matrix with entry (xa)u,v(x_a)_{u,v}(xa​)u,v​ at ((a,u),(a,v))((a,u),(a,v))((a,u),(a,v)) and zero between distinct blocks, and let PPP retain precisely those same-block entries of any matrix indexed by {0,…,n−1}×{0,…,m−1}\{0,\ldots,n-1\}\times\{0,\ldots,m-1\}{0,…,n−1}×{0,…,m−1}. The theorem asserts that JJJ is injective and

Sp(Jx)=(∑a=0n−1Sp(xa)p)1/pS_p(Jx)=\left(\sum_{a=0}^{n-1}S_p(x_a)^p\right)^{1/p}Sp​(Jx)=(a=0∑n−1​Sp​(xa​)p)1/p

for every such family xxx. In addition, if p≥1p\geq1p≥1, then P∘P=PP\circ P=PP∘P=P, the ranges of PPP and JJJ coincide, and Sp(PX)≤Sp(X)S_p(PX)\leq S_p(X)Sp​(PX)≤Sp​(X) for every matrix XXX of the indicated size. For 0<p<10<p<10<p<1, this last conjunction is only asserted conditionally on 1≤p1\leq p1≤p.

Human review
  • Endorsed by Shuze Chen · Oct 21, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 21, 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