Skip to content

Quantum Relative Entropy

The asymmetric quantum divergence Tr[ρ(logρ−logσ)], finite under support inclusion and monotone under quantum channels, comparing an ordered pair of density operators.

Version
v1 · 2026-09-28 · History
Domain-specific #
11617
Domain group
Formal Sciences
Origin domain
Information Theory
Subdomain
Quantum Information Theory → Information Theory
Aliases
Umegaki Relative Entropy, Quantum Kullback Leibler Divergence

Core Idea

Quantum relative entropy compares a state ρ with a reference state σ through the trace of ρ times the difference of their operator logarithms. It reduces to classical Kullback–Leibler divergence when the states commute in a common eigenbasis.

The argument order and support convention are essential: if ρ occupies a direction to which σ assigns zero support, the value is infinite. It is not a metric, but its nonnegativity and data-processing behavior make it central to quantum information, statistics, and thermodynamics.

Scope of Application

  • Quantum information. Bounds discrimination and coding tasks.
  • Quantum statistics. Compares states and hypotheses.
  • Operator algebras. Studies monotone divergences.
  • Quantum thermodynamics. Expresses free-energy and irreversibility relations.

Clarity

Specify ordered arguments, normalization, Hilbert-space setting, log base, spectral functional calculus, support relation, finite/infinite convention, and the channel or optimization under discussion. Inclusion test: Require ordered density operators, functional calculus for their logs, the support convention, and the ρ-weighted trace; state logarithm base and dimension assumptions. Exclusion test: Exclude von Neumann entropy of one state, trace distance or fidelity, symmetric metric claims, and classical KL applied to outcomes without specifying measurement. Nearest boundary: For commuting ρ and σ, simultaneous diagonalization reduces quantum relative entropy to classical KL; noncommuting states retain operator ordering and cannot be reduced to arbitrary eigenvalue pairing. Exit condition: Changing the functional, symmetrizing it, or ignoring the support condition yields another divergence rather than Umegaki quantum relative entropy. Common misclassifications: It is not the von Neumann entropy of one state. It is not symmetric and not a distance metric. Infinite value does not imply orthogonal states. Eigenvalues cannot be paired arbitrarily for noncommuting states. Nearest named distinctions: Von Neumann entropy: Is a one-state uncertainty functional. Trace distance: Is symmetric and metric-like. Quantum fidelity: Is a bounded overlap measure. Classical KL divergence: Matches only in an appropriate commuting representation.

Manages Complexity

One extended-real functional combines spectral calculus, support geometry, asymmetry, and information monotonicity while retaining the classical divergence in commuting cases.

Abstract Reasoning

  1. Validate the two density operators.
  2. Check support inclusion.
  3. Compute operator logarithms spectrally.
  4. Take the ρ-weighted trace difference.
  5. Interpret the ordered value through a justified information-theoretic inequality.

Knowledge Transfer

Classical relative-entropy intuition transfers exactly only for commuting states; in general, operator order, support, allowable measurements, and quantum-channel monotonicity must be preserved.

Relationships to Other Abstractions

Local relationship map for Quantum Relative EntropyParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.QuantumRelative EntropyDOMAINPrime abstraction: Comparison — is a kind ofComparisonPRIME

Current abstraction Quantum Relative Entropy Domain-specific

Parents (1) — more general patterns this builds on

  • Quantum Relative Entropy is a kind of Comparison Prime

    Quantum Relative Entropy is a strict kind of Comparison: it places an ordered pair of density operators in a shared operator frame and assigns their asymmetric divergence.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Quantum Relative Entropy sits in a moderately populated region (43rd percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Quantum States & Computational Models (12 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08