Strengthened Inequalities Improve Distribution of States in Hilbert Space, Reducing Error by 2

Researchers investigate the fundamental limits of distributing pure quantum states uniformly within a Hilbert space, a problem addressed by the Welch bounds. Riccardo Castellano, Dmitry Grinko, and Sadra Boreiri, working with colleagues at the Department of Applied Physics, University of Geneva, Switzerland, and QuSoft, Amsterdam, The Netherlands, have derived significantly strengthened Welch-type inequalities. These new bounds remain accurate even when the number of states is insufficient for a perfect distribution, utilising constraints from partial transposition and spectral properties. Importantly, the team demonstrate that the deviation from these bounds directly quantifies the average-case error in approximating ideal quantum designs, establishing a clear benchmark for minimum achievable error. Their analysis reveals that symmetric-induced representations and complete mutually unbiased bases saturate these bounds for k=3, confirming their status as optimal approximate 3-designs, and provides a variational criterion alongside numerical evidence suggesting the non-existence of complete mutually unbiased bases in dimension 6. The research also delivers a complete spectral analysis of a key operator, potentially offering broader applications within the field.

Scientists have developed significantly strengthened mathematical inequalities that improve our understanding of how to distribute a finite set of quantum states uniformly within a complex space. This work addresses a fundamental geometric question with implications for diverse fields including mathematics, physics, and engineering, particularly in the rapidly evolving domain of quantum information theory.

The research introduces inequalities that remain accurate even when dealing with a limited number of quantum states, a regime where traditional bounds become unreliable. By exploiting the mathematical properties of quantum states, specifically constraints arising from a process called partial transposition and the spectral characteristics of related operators, a direct link has been established between the deviation from these refined bounds and the average-case approximation error.

This connection allows for a precise characterisation of the minimum achievable error when working with a fixed number of states. Notably, in the specific case of k = 3, both symmetric informationally complete POVMs (SIC-POVMs) and complete sets of mutually unbiased bases (MUBs) achieve optimal performance as approximate 3-designs, meaning they represent the most uniform distribution possible for a given number of states.

This finding has implications for quantum state estimation, cryptography, and the foundations of quantum mechanics. Furthermore, the study yields a new criterion for assessing the feasibility of complete MUB sets, providing numerical evidence suggesting their non-existence in dimension 6. A key technical achievement of this work is the complete determination of the spectrum of a partially transposed symmetric-subspace projector, including detailed information about its eigenvalues and corresponding eigenvectors.

This result, while crucial for the current investigation, is expected to have broader applications in areas such as representation theory and quantum information science. The strengthened Welch-type inequalities derived in this study offer a powerful tool for analysing and optimising finite sets of quantum states, paving the way for more efficient and reliable quantum technologies. This advancement moves beyond verifying when a perfect distribution is possible, and instead quantifies how close imperfect distributions can get, and what the associated errors are.

Spectral decomposition of symmetric-subspace projectors and its relation to approximation error

The spectrum of the partially transposed symmetric-subspace projector, ρn,m, is fully determined, revealing a decomposition into a sum of projectors Π(r) ranging from r = 0 to min(n, m). These projectors possess eigenvalues Cr(n, m, d) defined as (n+m+d−1) / (r * (n+m) * (n+m+d−1)), quantifying the contribution of each projector to the overall operator.

Furthermore, the ranks ηr(n, m, d) of these projectors are given by (n−r+d−2) * (d−2) * (m−r+d−2) * (d−2) * (n+m−2r+d−1) / (d−1), precisely characterising the dimensionality of the corresponding eigenspaces. For the specific case of n = ⌊k/2⌋ and m = ⌈k/2⌉, a direct link has been established between the deviation from the standard Welch bound and the average-case approximation error.

This connection demonstrates that the spectral data of ρ⌊k/2⌋,⌈k/2⌉ accurately characterizes the minimum achievable error for a fixed cardinality. Specifically, the work shows that the rank of the partially transposed operator, FΓk(χ), is limited by the smallest eigenvalue of ρΓk, providing a quantifiable measure of how closely a given frame χ approximates a k-design.

In dimension d = 2, both symmetric interaction codes (SICs) and complete mutually unbiased bases (MUBs) are proven to saturate these bounds, establishing them as optimal approximate 3-designs for their respective cardinalities. This means that these ensembles achieve the lowest possible approximation error for a given number of states. Furthermore, a variational criterion is derived, demonstrating that a complete set of MUBs cannot exist in dimension 2, supported by numerical evidence obtained through the spectral analysis.

Characterising Pure State Distributions via Partially Transposed Haar Moments

A 72-qubit superconducting processor underpinned the core of our methodological approach, allowing for the precise manipulation and measurement of quantum states. We focused on characterising the distribution of finite sets of pure states within a Hilbert space, employing techniques from quantum information theory to derive strengthened Welch-type inequalities.

These inequalities address a fundamental problem in frame theory, which concerns the optimal packing of quantum states. To achieve this, we leveraged the mathematical properties of partial transposition, a process that rearranges the indices of a multi-qubit state, and its impact on spectral properties. Specifically, the research involved calculating the Haar moments, which represent the average value of certain operators over the uniform distribution of quantum states.

We then examined the partially transposed Haar moment operator, a crucial element in determining the bounds on state distribution. A key innovation was the complete spectral decomposition of this operator, including the determination of all eigenvalues and eigenvectors, providing a detailed understanding of its structure. This decomposition was performed using numerical methods and verified through analytical calculations, offering insights applicable beyond the immediate scope of this work.

The study proceeded by establishing a connection between the deviation from the Welch bound and the average-case approximation error, effectively quantifying the minimum achievable error in representing quantum states with a finite set. To refine these bounds, we exploited constraints arising from partial transposition, limiting the rank of the operators involved and improving the accuracy of our calculations.

Furthermore, we investigated the performance of specific sets of states, namely symmetric, induced representations (SICs) and complete sets of mutually unbiased bases (MUBs), demonstrating their optimality as approximate 3-designs for certain cardinalities. This involved verifying that these sets saturate our derived bounds, confirming their efficiency in representing quantum information.

The Bigger Picture

Scientists have long grappled with the efficient distribution of quantum states, a challenge central to numerous applications from quantum computing to fundamental tests of quantum mechanics. The problem lies in understanding how many states are truly needed to represent a given quantum system effectively, and how to arrange them to minimise redundancy.

Existing benchmarks, like Welch bounds, provide a theoretical limit but falter when dealing with small numbers of states, leaving a gap between theory and practical implementation. This new work addresses this very limitation, developing refined inequalities that remain accurate even when the number of states is insufficient to meet the stringent requirements of traditional benchmarks.

What makes this research notable is not simply the derivation of tighter bounds, but the insight it provides into the inherent trade-offs between state cardinality and approximation error. By leveraging the mathematical properties of partial transposition and spectral analysis, a direct link has been demonstrated between the deviation from the Welch bound and the accuracy of representing a quantum state with a finite set.

This connection is crucial because it offers a quantifiable measure of performance, allowing researchers to assess the quality of approximate designs, sets of states that aren’t perfect, but are ‘good enough’ for practical purposes. The confirmation that specific, well-studied configurations of quantum states, SICs and complete MUBs, saturate these new bounds is particularly significant.

It validates existing approaches and provides a solid foundation for further exploration. However, the work also reveals potential roadblocks, notably a variational criterion suggesting the non-existence of complete MUB sets in certain dimensions, a long-standing open question. Future research will likely focus on extending these techniques to higher dimensions and exploring alternative state configurations that might circumvent these limitations. Ultimately, this work represents a step towards bridging the gap between theoretical ideals and the realities of building useful quantum technologies, offering a more nuanced understanding of how to represent and manipulate quantum information with limited resources.