⚛️ quantum physics

Quantum Optimal Transport Barycenters: Existence, Duality, and Gaussian Rigidity

This paper establishes a comprehensive Quantum Optimal Transport barycenter framework with existence and duality results, proving that Gaussian input states (specifically when at least one is faithful) uniquely determine a Gaussian barycenter by bridging the gap between covariance uniqueness and full quantum state uniqueness through a novel state-reconstruction principle.

Original authors: Augusto Gerolin, Zhiyi Lin

Published 2026-10-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Augusto Gerolin, Zhiyi Lin

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quantum world, where particles exist in states of probability rather than fixed positions, scientists have long sought a way to measure the distance between two different configurations of matter. Just as a geographer might calculate the most efficient route to move a pile of sand from one location to another, physicists use a concept called optimal transport to find the most efficient way to transform one quantum state into another. This mathematical framework, known as the Wasserstein distance, has become a standard tool for understanding how quantum systems evolve and interact. However, a more complex challenge arises when researchers need to find a "middle ground" between not just two, but many different quantum states simultaneously. In classical physics, this middle point, or barycenter, is well understood: if you average several Gaussian distributions, the result is always another Gaussian distribution, and it is unique. The question that has lingered in the quantum realm is whether this same simplicity holds true when the objects being averaged are quantum states, which are far more fragile and complex than classical clouds of probability.

A team of researchers has now answered this question with a definitive framework that unifies two different ways of describing quantum transport. They developed a method to find the quantum barycenter for a broad class of systems, proving that such a middle state always exists and can be found using a specific mathematical duality. Their work bridges the gap between two previously separate approaches: one that treats the transport as a relationship between two quantum states, and another that treats it as a transformation or channel acting on a state. By treating both scenarios under a single mathematical roof, the authors established that the problem of finding the average quantum state is solvable and that the solution obeys strict rules of energy and structure.

The most significant discovery in this work concerns the nature of the solution when the starting states are Gaussian. In quantum mechanics, a Gaussian state is a specific type of configuration that is mathematically well-behaved and often used to model light and other continuous variables. The researchers proved that if you start with a collection of Gaussian states, the resulting average will also be a Gaussian state. This might seem like a natural extension of classical physics, but in the quantum world, it is not guaranteed. Quantum states can share the same basic statistical properties, such as average position and spread, yet still be fundamentally different from one another. The authors showed that while the optimal covariance matrix determining these statistical properties is unique, this uniqueness does not, in general, guarantee that the underlying quantum state itself is unique.

However, the team identified a precise condition that forces the solution to be unique. They found that if at least one of the input states is "faithful"—meaning it has a non-zero probability of being found in any possible configuration within its system—then the resulting average is not only Gaussian but is the one and only possible quantum state that satisfies the conditions. This finding resolves a long-standing ambiguity. Without this condition, the researchers demonstrated that multiple different quantum states, including some that are not Gaussian at all, could all serve as valid averages for the same set of inputs. This means that simply knowing the average position and spread of the inputs is not enough to determine the final quantum state unless the inputs are sufficiently "full" or faithful.

The paper also explores what happens when the quantum effects become very small, approaching the scale of everyday classical physics. The researchers proved that as the fundamental quantum scale shrinks, the quantum barycenter smoothly transitions into the classical average. The shape and position of the quantum solution converge exactly to the classical Gaussian average, confirming that the new framework is consistent with the laws of physics we observe in the macroscopic world. Furthermore, they applied their theory to a specific and common type of quantum state known as a thermal state, which describes systems in equilibrium with their environment. For these states, they derived an explicit formula that allows scientists to calculate the exact average without needing to solve the complex optimization problem from scratch.

This work provides a complete and rigorous map for navigating the landscape of quantum averages. It confirms that while the quantum world allows for surprising complexities where different states can look identical in their basic statistics, there are clear boundaries where uniqueness returns. By establishing that a single faithful input is sufficient to pin down a unique, Gaussian solution, the authors have provided a powerful tool for quantum information science. Their results ensure that when scientists need to average quantum data or find a central state for a network of quantum systems, they can rely on a framework that is both mathematically sound and physically precise, bridging the gap between abstract theory and the concrete behavior of quantum matter.

Technical Summary: Quantum Optimal Transport Barycenters

1. Problem Statement

This paper establishes a rigorous mathematical framework for Quantum Optimal Transport (QOT) barycenters, serving as the quantum analog to the classical Wasserstein barycenters for probability measures [1]. The central problem is defined as finding a quantum state $\rho$ that minimizes a weighted sum of transport costs to a given family of input states $\sigma_s$ :

$\text{BC}_{C,\alpha}^{A,E}(\sigma) := \inf_{\rho \in S_{A,E}(H_0)} \sum_{s=1}^N \alpha_s \text{QOT}(\rho, \sigma_s \mid C_s)$

where $C_s$ are nonnegative self-adjoint cost operators, $\alpha_s$ are positive weights summing to 1, and the search space $S_{A,E}(H_0)$ may include energy constraints.

The paper specifically targets two prominent static transport models:

Quantum-State Formulation ( $\tau = +1$ ): Based on the 2-quantum Wasserstein distance introduced by Caglioti–Golse–Mouhot–Paul [7, 22], where transport costs are defined between density operators.
Quantum-Channel Formulation ( $\tau = -1$ ): Based on the work of De Palma and Trevisan [14], where transport plans are interpreted as quantum channels, corresponding to couplings between a state and the transpose of another.

A primary motivation is to understand the Gaussian rigidity of these barycenters: Do Gaussian input states force the optimal barycenter to be Gaussian and unique among all quantum states, or only within the Gaussian family? This is non-trivial because, unlike classical measures, distinct quantum states can share identical first and second moments (mean and covariance).

2. Methodology

The authors employ a variational approach on separable Hilbert spaces, utilizing operator-theoretic tools and convex duality.

Existence and Compactness: The existence of barycenters is proven via the direct method of the calculus of variations. To handle infinite-dimensional spaces where marginal constraints alone do not ensure compactness, the authors impose conditions on the cost operators: either they must have a compact resolvent (covering confining Hamiltonians) or be quadrature-generated (quadratic costs based on canonical observables).
Duality: A strong duality result is established using the Fenchel–Rockafellar theorem. The primal problem (minimizing transport cost over couplings) is dualized into a maximization problem over bounded self-adjoint operator potentials. This framework unifies the state and channel formulations.
Gaussian Reduction: For quadratic costs, the transport cost depends only on the first and second canonical moments. The authors utilize Gaussian projection (mapping a state to the unique Gaussian state with the same mean and covariance) to show that the infinite-dimensional barycenter problem reduces to a finite-dimensional convex optimization problem over covariance matrices.
State Reconstruction and Rigidity: To bridge the gap between the uniqueness of the optimal covariance and the uniqueness of the full quantum state, the authors develop a reconstruction principle. They analyze the covariance slackness relation derived from the dual problem. By showing that a faithful input (a state with full-rank covariance) imposes a specific structure on the optimal coupling, they prove that the coupling must be Gaussian, thereby forcing the barycenter to be Gaussian and unique.
Semiclassical Analysis: The paper analyzes the limit as Planck's constant $\hbar \to 0^+$ , demonstrating convergence to the classical Gaussian 2-Wasserstein barycenter.

3. Key Contributions and Results

A. General Existence and Duality (Theorems 2.1–2.2)

The paper proves the existence of QOT barycenters for a broad class of costs on separable Hilbert spaces, including unbounded costs. It establishes a Kantorovich-type duality formula with energy constraints, valid for both the quantum-state and quantum-channel formulations.

B. Gaussian Reduction (Theorem 2.5)

For inputs that are Gaussian states, the authors prove that:

A Gaussian barycenter always exists.
The problem reduces to a finite-dimensional convex optimization over covariance matrices (Equation 2.11).
The mean of the barycenter is the weighted average of the input means.

C. Global Uniqueness and Gaussian Rigidity (Theorem 2.6)

This is the paper's central result. It establishes that if the input states are Gaussian and at least one input is faithful (i.e., its covariance matrix satisfies $\Sigma + i\frac{\hbar}{2}\Omega \succ 0$ ), then:

The barycenter is unique among all quantum states (not just Gaussian ones).
The unique barycenter is necessarily Gaussian.
The proof relies on a reconstruction principle: the covariance slackness relation, combined with the faithfulness of one input, forces the optimal coupling to be Gaussian, which in turn forces the barycentric marginal to be Gaussian.

D. Semiclassical Convergence (Theorem 2.7)

The authors prove that as $\hbar \to 0^+$ , the quantum barycenter covariances converge to the covariance of the classical Gaussian 2-Wasserstein barycenter. The associated Wigner distributions converge weakly to the classical Gaussian measure.

E. Explicit Characterizations (Corollary 2.8)

For displaced isotropic thermal states (a specific class of Gaussian states), the paper provides explicit formulas for the unique barycenter in both the quantum-state and quantum-channel formulations.

F. Non-Uniqueness on the Boundary (Example 5.31)

The paper demonstrates that the faithfulness condition is sufficient but not necessary. If all inputs are partially pure (some modes are pure, others mixed), uniqueness can fail. The authors construct an example where distinct Gaussian and non-Gaussian barycenters exist with identical first and second moments, highlighting that without a faithful input, the optimal covariance does not uniquely determine the quantum state.

4. Significance and Claims

The paper claims to provide a unified treatment of 2-quantum Wasserstein barycenters, resolving the tension between the quantum-state and quantum-channel formulations under a single theoretical umbrella.

Its primary significance lies in the Gaussian rigidity theorem. While classical optimal transport guarantees that Gaussian inputs yield a unique Gaussian barycenter, the quantum setting introduces ambiguity because moments do not uniquely determine a state. This work proves that faithfulness is the precise condition required to restore this classical-like rigidity in the quantum domain.

The authors modestly note that while faithfulness is sufficient for uniqueness, it is not necessary (as seen in the case of pure inputs), but its absence on the "partially pure boundary" can lead to non-uniqueness. The framework is presented as a variational theory for static transport, distinct from dynamical approaches (like Carlen–Maas), and is intended to serve as a foundational tool for analyzing quantum transport problems involving quadratic costs.