[Paper Review] Maximum Geometric Quantum Entropy

Fabio Anzà, James P. Crutchfield|arXiv (Cornell University)|Aug 19, 2020

Quantum Information and Cryptography29 references4 citations

TL;DR

This paper introduces a maximum entropy principle to estimate geometric quantum states from a given density matrix, using a geometrically motivated entropy functional. The key result is a closed-form Gaussian-like distribution on the pure-state manifold, parameterized by the density matrix, which provides a physically richer state description than the density matrix alone.

ABSTRACT

Any given density matrix can be represented as an infinite number of ensembles of pure states. This leads to the natural question of how to uniquely select one out of the many, apparently equally suitable, possibilities. Following Jaynes' information-theoretic perspective, this can be framed as an inference problem. We propose the Maximum Geometric Quantum Entropy Principle to exploit the notions of Quantum Information Dimension and Geometric Quantum Entropy. These allow us to quantify the entropy of fully arbitrary ensembles and select the one that maximizes it. After formulating the principle mathematically, we give the analytical solution to the maximization problem in a number of cases and discuss the physical mechanism behind the emergence of such maximum entropy ensembles.

Motivation & Objective

• To address the limitation of density matrices in fully capturing the physical realization of quantum ensembles.
• To develop a principled method for estimating geometric quantum states—probability distributions on the pure-state manifold—given only a density matrix.
• To bridge the gap between operational quantum tomography (yielding density matrices) and physically meaningful state descriptions via geometric quantum mechanics.
• To establish a maximum entropy framework tailored to the geometric structure of quantum state space, ensuring consistency with quantum measurement statistics.

Proposed method

• Formalizes geometric quantum states as probability measures on the complex projective space $\mathbb{C}P^{D-1}$, the manifold of pure states.
• Adopts the Fubini-Study volume element $dV_{FS}$ as the invariant measure for defining geometric quantum entropy.
• Applies the maximum entropy principle using the geometric quantum entropy functional, constrained by the known density matrix $\rho$.
• Derives the maximum entropy geometric state as a multivariate Gaussian distribution on the pure-state manifold, parameterized by $\vec{\mu}$ and $\Sigma$, with $\Sigma = \rho$ when $\vec{\mu} = 0$.
• Expresses the resulting estimate as $q_{\text{ME}}(Z) = Q^{-1} \exp\left(-\frac{1}{2} \vec{Z}^* \rho^{-1} \vec{Z}\right)$, where $Q$ is the normalization integral over $\mathcal{P}(\mathcal{H})$.
• Uses the relation $\rho_{\alpha\beta} = \mathbb{E}[Z^\alpha \overline{Z}^\beta]$ to connect the density matrix to the moments of the geometric state.

Experimental results

Research questions

• RQ1 How can one estimate a geometric quantum state from a given density matrix, given that the density matrix alone does not fully specify the physical ensemble?
• RQ2 What is the appropriate entropy functional for quantum state estimation in the geometric quantum mechanics framework?
• RQ3 Can a maximum entropy principle be applied to geometric quantum states to yield a unique, physically meaningful estimate?
• RQ4 What is the functional form of the maximum entropy geometric state when constrained by a known density matrix?

Key findings

• The maximum entropy geometric quantum state is given by $q_{\text{ME}}(Z) = Q^{-1} \exp\left(-\frac{1}{2} \vec{Z}^* \rho^{-1} \vec{Z}\right)$, with $Q = \int_{\mathcal{P}(\mathcal{H})} dV_{FS} \, \exp\left(-\frac{1}{2} \vec{Z}^* \rho^{-1} \vec{Z}\right)$, when $\det \rho \neq 0$.
• The estimate reduces to a multivariate Gaussian distribution on the pure-state manifold, with covariance matrix $\Sigma = \rho$ and zero mean $\vec{\mu} = 0$ when only the density matrix is known.
• The geometric quantum state provides a more informative description of a quantum system than the density matrix alone, as multiple geometric states can yield the same $\rho$.
• The method ensures consistency with all POVM statistics, as the density matrix $\rho_q$ derived from $q(Z)$ matches the input $\rho$ via $\rho_{\alpha\beta} = \mathbb{E}[Z^\alpha \overline{Z}^\beta]$.
• The approach is invariant under changes of basis and respects the underlying differential geometry of the pure-state manifold.
• An example with a single qubit ($\rho_{00} = 0.55$, $\rho_{11} = 0.45$, $\rho_{01} = 0.2 - 0.3i$) demonstrates the method's feasibility and visualizes the resulting state in $(p, \phi)$ coordinates.