Alphabeta Math
DefinitionDefinition: AI-adaptedProof: AI-adaptedPipeline-generated
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Quantum spectral entropy and relative-entropy domains

Definition

For a trace-class density ρ on a Hilbert space H, define its support subspace in either finite or infinite dimension by supp⁡ρ:=(ker⁡ρ)⊥,ker⁡ρ:={x∈H:ρx=0}. It is closed; write Pρ for its orthogonal projection, whose existence and uniqueness follow from Orthogonal decomposition by a closed subspace.

Fix the positive Boltzmann scale kB>0 (in SI units J/K, as in the approved QSM quantities design and Trace-state cyclicity and spectral probability calculus). For a trace-class density ρ with positive eigenvalues pi, define S(ρ)=−kB∑ipilog⁡pi∈[0,∞], with 0log⁡0=0. The logarithm is natural and acts on dimensionless density eigenvalues.

For finite-dimensional densities ρ,σ, define supp⁡ρ=(ker⁡ρ)⊥. If supp⁡ρ⊈supp⁡σ, set D(ρ∥σ)=+∞. Otherwise define D(ρ∥σ)=Tr⁡(ρlog⁡ρ)−Tr⁡(ρlog⁡σ), where each logarithm is restricted to the positive support of its operator and zero eigenvalues contribute zero to Tr⁡(ρlog⁡ρ).

For trace-class normal densities on a separable Hilbert space, write ρ=∑ipi∣ui⟩⟨ui∣ and σ=∑jsj∣vj⟩⟨vj∣ on their positive supports. Define the nonnegative cross-log moment C(ρ,σ):=∑i,jpi(−log⁡sj)∣⟨ui,vj⟩∣2∈[0,∞]. This is the spectral first moment of −log⁡σ in the scalar state law of ρ; equivalently it is the nonnegative sum over the spectral projections, so it is independent of bases within eigenspaces. When support inclusion holds and C(ρ,σ) and S(ρ) are both finite, define D(ρ∥σ)=C(ρ,σ)−S(ρ)/kB. If C=+∞ and S(ρ)<∞, set D=+∞. No value is assigned by this definition to +∞−(+∞); Trace-class cross-log entropy inequality proves that finite C forces finite S(ρ).

The spectral-moment identification in the preceding cross-log clause is restricted to supp⁡ρ⊆supp⁡σ, so that the state is supported on the space where −log⁡σ acts. Outside that domain, C denotes only the displayed nonnegative double sum.

Facts & Assumptions

Given: The finite spectral calculus, trace-class normal density states, the scalar PVM law on a nonzero separable complex Hilbert space, and the positive scale kB>0.

[A1]

The Axiom of Choice is inherited through the trace-state spectral-calculus supplier closure; the finite-dimensional clauses themselves are choice-free (The Axiom of Choice).

[F1]

Finite matrix spectral and state calculus: finite Hermitian matrices have real eigenvalues and logarithms are defined on their positive supports.

[F2]

Trace-class normal density states and Trace-state cyclicity and spectral probability calculus: trace-class densities have countable positive spectral weights and von Neumann entropy is the nonnegative spectral sum, possibly infinite.

Proof

1.1givenA1F1algebra

In finite dimension, support inclusion means every positive eigenvector of ρ lies in the span of positive eigenvectors of σ. Each positive eigenvalue in either support is a positive real number, so its logarithm is finite; the zero eigenspaces contribute no ρ-weight under the stated support inclusion. Therefore both trace terms in the finite formula are finite real sums, and the formula has no undefined product 0log⁡0. The Axiom of Choice is present only through the explicitly retained spectral-calculus supplier closure.

2.1givenF2algebra∎

In the trace-class case, every summand in C(ρ,σ) is nonnegative because 0<sj≤1. The sum is therefore defined unambiguously as the supremum of its finite partial sums in [0,∞]. Grouping equal eigenvalues replaces the basis coefficients by traces of products of the corresponding spectral projections, so the resulting value is independent of basis choices within eigenspaces. The definition distinguishes finite subtraction from the undefined case of two infinite terms.

Depends on

Dependency tree · 2 levels

Sources