Spectral expansion of a symmetric operator's quadratic form against its eigenbasis
ProvedHlawkaSchatten.re_inner_apply_eq_sum_eigenbasisLet be a complex inner-product space and let be a symmetric (self-adjoint) complex-linear map, meaning for all . Let be a finite index set, let be an orthonormal basis of , and let be such that each is an eigenvector of with eigenvalue , i.e. for every . Then, for every ,
This is the spectral expansion of the (real) quadratic form of a symmetric operator against its eigenbasis: writing in eigen-coordinates, decomposes into a weighted sum of the eigenvalues , weighted by the squared eigen-coordinates . It is the first step of the spectral-lift layer: applied to the Hermitian dilations of a pair of rectangular operators, it lets Bregman-type and Mazur-type trace quantities be rewritten as sums over eigenvalues weighted by basis overlaps.
Formalization Note. Having an orthonormal basis indexed by a finite set makes finite-dimensional as a consequence, not as a separately stated hypothesis; the theorem places no other restriction on , , or .
import Mathlib.Analysis.Calculus.LHopital
import Mathlib.Analysis.Convex.Deriv
import Mathlib.Analysis.Convex.SpecificFunctions.Basic
import Mathlib.Analysis.InnerProductSpace.Basic
import Mathlib.Analysis.InnerProductSpace.Dual
import Mathlib.Analysis.InnerProductSpace.NormPow
import Mathlib.Analysis.InnerProductSpace.PiL2
import Mathlib.Analysis.InnerProductSpace.Trace
import Mathlib.Data.Sign.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
-/
/-!
# Finite Hermitian spectral trace expansions
This file connects the overlap-weighted scalar comparison to traces of
finite-dimensional symmetric complex-linear maps.
-/
open scoped InnerProductSpace
open RCLike
open ComplexConjugate
variable {ι κ E : Type*} [Fintype ι] [Fintype κ]
[NormedAddCommGroup E] [InnerProductSpace ℂ E]
theorem HlawkaSchatten.re_inner_apply_eq_sum_eigenbasis
(A : E →ₗ[ℂ] E) (hA : A.IsSymmetric)
(e : OrthonormalBasis ι ℂ E) (a : ι → ℝ)
(he : ∀ i, A (e i) = (a i : ℂ) • e i) (x : E) :
(⟪x, A x⟫_ℂ).re = ∑ i, a i * ‖(⟪e i, x⟫_ℂ)‖ ^ 2 := by sorry