Quantum weighted entropy and its properties

Yuri Suhov, Stefan ZohrenarXiv 2014 · DOI · arXiv

Abstract

We introduce quantum weighted entropy in analogy to an earlier notion of (classical) weighted entropy and derive many of its properties. These include the subadditivity, concavity and strong subadditivity property of quantum weighted entropy, as well as an analog of the Araki-Lieb inequality. Interesting byproducts of the proofs are a weighted analog of Klein's inequality and non-negativity of quantum weighted relative entropy. A main difficulty is the fact that the weights in general do not commute with the density matrices.

Introduction

In standard information theory, Shannon entropy and its quantum counterpart, von Neumann entropy, serve as the foundational measures of uncertainty. However, these measures treat all outcomes or quantum states as equally significant in terms of their structural contribution to the total entropy; only their respective probabilities (or eigenvalues) dictate the result. In many physical and statistical contexts, certain states carry more "importance" or "weight" than others. While classical statistics has long utilized "weighted entropy" to account for these differences, a robust quantum mechanical generalization has remained elusive due to the mathematical complexities of non-commutative algebra.

This paper establishes a formal framework for quantum weighted entropy, extending the von Neumann entropy by incorporating a weight matrix. The authors derive essential properties—including subadditivity, concavity, and strong subadditivity—while addressing the fundamental challenge of non-commuting operators.

Defining Quantum Weighted Entropy

The authors define the quantum weighted entropy $S_\phi(\rho)$ as: $$S_\phi(\rho) = - \text{tr}(\phi \rho \log \rho)$$ where $\rho$ represents the density matrix of the quantum system and $\phi$ is a Hermitian, positive-definite weight matrix.

By introducing $\phi$, the entropy becomes sensitive to the specific characteristics of the quantum system beyond mere uncertainty. A key conceptual tool introduced is the $\rho$-conjugate weight. The paper demonstrates that different weight matrices can yield identical entropy values if they are effectively equivalent from the perspective of the state's eigenvectors, allowing for a more flexible treatment of weighting factors.

Mathematical Foundations: The Non-Commutativity Hurdle

The transition from classical to quantum weighted entropy is non-trivial because, in general, the weight matrix $\phi$ and the density matrix $\rho$ do not commute ($[\phi, \rho] \neq 0$). In classical information theory, weights are scalars, and the order of multiplication is irrelevant. In the quantum realm, this non-commutativity breaks standard proofs used for traditional entropy measures.

To overcome this, the authors first establish a Weighted Klein’s Inequality. This inequality serves as the primary mathematical engine for the paper, facilitating the proof of the Weighted Gibbs Inequality. A critical finding is that for weighted relative entropy to remain non-negative, a specific "trace condition" must be satisfied: $\text{tr}(\phi \rho) \ge \text{tr}(\phi \sigma)$. This suggests that the total "weighted energy" or significance of the states must be carefully balanced to maintain the physical consistency of the entropy measure.

Core Properties and Inequalities

The paper successfully extends several landmark inequalities of quantum information theory to the weighted case:

  • Subadditivity: The authors prove that the weighted entropy of a joint system is less than or equal to the sum of the weighted entropies of its individual components, provided specific trace conditions are met.
  • Concavity: Weighted entropy is shown to be a concave function. This implies that the process of mixing quantum states increases (or maintains) the overall weighted uncertainty of the system.
  • Araki-Lieb Inequality: The research derives a weighted version of this "triangle inequality," relating the entropy of a composite system to the difference in entropies of its constituent parts.
  • Strong Subadditivity (SSA): Often considered the most important inequality in quantum information theory, SSA is also addressed. The authors provide a proof for the weighted version, though it currently requires a "technical condition" where certain matrices must commute.

Limitations and Future Directions

The primary limitation of the current work is the reliance on commutation assumptions for the proof of Strong Subadditivity. Unlike the standard von Neumann entropy, where SSA holds universally, the weighted version's proof is not yet optimal. The authors suggest that future research should focus on relaxing these commutation requirements to achieve a more general theory.

Furthermore, while the mathematical framework is now robustly established, the physical applications—such as in quantum thermodynamics, coding theory, or many-body physics—remain to be explored. This theoretical foundation opens the door for researchers to apply weighted measures to systems where state significance is non-uniform.

Summary updated

Abstract

We introduce quantum weighted entropy in analogy to an earlier notion of (classical) weighted entropy and derive many of its properties. These include the subadditivity, concavity and strong subadditivity property of quantum weighted entropy, as well as an analog of the Araki-Lieb inequality. Interesting byproducts of the proofs are a weighted analog of Klein's inequality and non-negativity of quantum weighted relative entropy. A main difficulty is the fact that the weights in general do not commute with the density matrices.
BibTeX
@misc{suhov2014quantum,
  author = {Yuri Suhov and Stefan Zohren},
  title = {{Quantum weighted entropy and its properties}},
  year = {2014},
  eprint = {1411.0892},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1411.0892},
}

Loading related content...

Figure