Prove2Me
Navigate
DiscoverFormalpediaBlogsUsersMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

The rectangular Bregman objective attains its minimum on the Schatten power sphere

Proved
HlawkaSchatten.rectangularBregmanObjective_isGlobalMinimumValue

by savarin · Sep 27, 2026 · Mathlib 0df444a (Lean v4.33.1)

global-minimumhlawka-schattenschatten-normvariational-methods

Let E,FE,FE,F be finite-dimensional complex inner-product spaces, let ι\iotaι be a nonempty finite index type, let p∈Rp\in\mathbb Rp∈R with p>1p>1p>1, and let a:ι→Ra:\iota\to\mathbb Ra:ι→R (no sign condition on aaa). For a complex-linear map T:E→FT:E\to FT:E→F with singular values sv⁡k(T)\operatorname{sv}_k(T)svk​(T), the Schatten ppp-power sphere is {T∣∑ksv⁡k(T)p=1}\{T\mid\sum_k\operatorname{sv}_k(T)^p=1\}{T∣∑k​svk​(T)p=1} (schattenPowerSphere), and let u:ι→u:\iota\tou:ι→ (that sphere).

For T:E→FT:E\to FT:E→F, write T^\widehat TT for its Hermitian dilation, the self-adjoint operator on E⊕FE\oplus FE⊕F with T^(x,y)=(T∗y, Tx)\widehat T(x,y)=(T^\ast y,\,Tx)T(x,y)=(T∗y,Tx) (hermitianDilation). For a self-adjoint operator CCC and h:R→Rh:\mathbb R\to\mathbb Rh:R→R, write h(C)h(C)h(C) for the result of applying hhh to CCC's eigenvalues in any orthonormal eigenbasis of CCC; this operator itself does not depend on which eigenbasis is chosen, since it acts as h(μ)h(\mu)h(μ) on every μ\muμ-eigenvector of CCC, not only on the vectors of one particular chosen eigenbasis — which is what lets S^∘Gp(T^)\widehat S\circ G_p(\widehat T)S∘Gp​(T) and T^∘Gp(T^)\widehat T\circ G_p(\widehat T)T∘Gp​(T) below be single well-defined operators. For S,T:E→FS,T:E\to FS,T:E→F, the dilated Bregman trace is

Bp(S,T)  =  Tr⁡Fp(S^)−Tr⁡Fp(T^)−Re⁡Tr⁡ ⁣(S^∘Gp(T^))+Tr⁡ ⁣(T^∘Gp(T^))B_p(S,T) \;=\; \operatorname{Tr}F_p(\widehat S) - \operatorname{Tr}F_p(\widehat T) - \operatorname{Re}\operatorname{Tr}\!\big(\widehat S\circ G_p(\widehat T)\big) + \operatorname{Tr}\!\big(\widehat T\circ G_p(\widehat T)\big)Bp​(S,T)=TrFp​(S)−TrFp​(T)−ReTr(S∘Gp​(T))+Tr(T∘Gp​(T))

(dilatedBregmanTrace), with Fp(x)=∣x∣p/pF_p(x)=|x|^p/pFp​(x)=∣x∣p/p and Gp(x)=∣x∣p−2xG_p(x)=|x|^{p-2}xGp​(x)=∣x∣p−2x. The rectangular Bregman objective is the half-dilated weighted sum

rectangularBregmanObjective(p,a,u,v)  =  ∑i∈ιai⋅Bp(ui,v)2,\mathrm{rectangularBregmanObjective}(p,a,u,v) \;=\; \sum_{i\in\iota} a_i\cdot\frac{B_p(u_i,v)}{2},rectangularBregmanObjective(p,a,u,v)=i∈ι∑​ai​⋅2Bp​(ui​,v)​,

as a function of vvv ranging over the Schatten ppp-power sphere, and rectangularWeightedSum(a,u)=∑iai ui\mathrm{rectangularWeightedSum}(a,u)=\sum_i a_i\,u_irectangularWeightedSum(a,u)=∑i​ai​ui​ is the (unnormalized) weighted sum of the underlying maps. Writing ∥T∥p=(∑ksv⁡k(T)p)1/p\|T\|_p=\big(\sum_k\operatorname{sv}_k(T)^p\big)^{1/p}∥T∥p​=(∑k​svk​(T)p)1/p for the Schatten ppp-norm (schattenPNorm), the theorem states that rectangularBregmanObjective(p,a,u,⋅)\mathrm{rectangularBregmanObjective}(p,a,u,\cdot)rectangularBregmanObjective(p,a,u,⋅) attains the global minimum value

(∑i∈ιai)−∥rectangularWeightedSum(a,u)∥p\Big(\sum_{i\in\iota} a_i\Big) - \big\|\mathrm{rectangularWeightedSum}(a,u)\big\|_p(i∈ι∑​ai​)−​rectangularWeightedSum(a,u)​p​

over the Schatten ppp-power sphere.

This is the operator-level analogue, for the Schatten ppp-norm on rectangular maps, of the exact Hilbert-space minimum of a weighted squared-distance objective: it is one of the two variational facts (the other being the analogous minimum for the paired Mazur-distance objective) whose comparison produces the two-sided bound between triple and pair deficits that this construction needs.

Formalization Note The Schatten ppp-power sphere is defined by ∑ksv⁡k(T)p=1\sum_k\operatorname{sv}_k(T)^p=1∑k​svk​(T)p=1; for p>0p>0p>0 this is equivalent to ∥T∥p=1\|T\|_p=1∥T∥p​=1 (schattenPNorm_eq_one_iff). In this theorem's range p>1p>1p>1, FpF_pFp​ is convex and differentiable with derivative GpG_pGp​. The proof uses the resulting scalar Bregman nonnegativity through a supporting theorem, then lifts it to BpB_pBp​ by the spectral trace identity.

Preamble
import Definitions.Def_HlawkaSchatten_SchattenNorm
import Definitions.Def_HlawkaSchatten_Variational
import Mathlib.Analysis.Calculus.LHopital
import Mathlib.Analysis.Convex.Deriv
import Mathlib.Analysis.Convex.SpecificFunctions.Basic
import Mathlib.Analysis.InnerProductSpace.Adjoint
import Mathlib.Analysis.InnerProductSpace.Basic
import Mathlib.Analysis.InnerProductSpace.Dual
import Mathlib.Analysis.InnerProductSpace.NormPow
import Mathlib.Analysis.InnerProductSpace.PiL2
import Mathlib.Analysis.InnerProductSpace.ProdL2
import Mathlib.Analysis.InnerProductSpace.SingularValues
import Mathlib.Analysis.InnerProductSpace.Trace
import Mathlib.Data.Sign.Basic
import Mathlib.LinearAlgebra.Matrix.Charpoly.Basic
import Mathlib.Topology.Compactification.OnePoint.Basic
import Mathlib.Topology.Instances.Sign

/-
Copyright (c) 2026 Ezzeri Esa. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Ezzeri Esa
-/

/-!
# Variational minima for the Bregman--Mazur argument

This file proves two reusable parts of the variational layer.  First, a
pointwise two-sided comparison transports to attained global minima, even
when the two objectives are indexed by different but equivalent spheres.
Second, the weighted squared-distance objective on a Hilbert unit sphere has
the exact minimum used in the Schatten argument.
-/


open scoped InnerProductSpace ComplexConjugate







variable {𝕜 H ι : Type*} [RCLike 𝕜] [Fintype ι]
  [NormedAddCommGroup H] [InnerProductSpace 𝕜 H]











section RectangularMazur

variable {E F κ : Type*} [Fintype κ]
  [NormedAddCommGroup E] [InnerProductSpace ℂ E] [FiniteDimensional ℂ E]
  [NormedAddCommGroup F] [InnerProductSpace ℂ F] [FiniteDimensional ℂ F]

open HlawkaSchatten
Formal statement
theorem HlawkaSchatten.rectangularBregmanObjective_isGlobalMinimumValue
    [Nonempty ι] {p : ℝ} (hp : 1 < p) (a : ι → ℝ)
    (u : ι → schattenPowerSphere (𝕜 := ℂ) (E := E) (F := F) p) :
    IsGlobalMinimumValue (rectangularBregmanObjective p a u)
      ((∑ i, a i) - schattenPNorm p (rectangularWeightedSum a u)) := by sorry
Source
https://github.com/savarin/hlawka-schatten/blob/79aa498bfcf7b22bd91d771fb32ec278e2d4704b/HlawkaSchatten/Variational.lean#L233-L310

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 paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me