Experimental High-Dimensional Quantum Overlapping Tomography

arXiv:2608.15795 · quant-ph · Submitted 2026-08-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: We're looking at the paper: "Experimental High-Dimensional Quantum Overlapping Tomography".

Title and authors: Kai: So, to wrap up on the main findings of "Experimental High-Dimensional Quantum Overlapping Tomography," they show that using graph theory for these high dimensional QOT allows them to reconstruct all six two body marginals from only twenty five measurement settings on a four party state.

Mira: Yeah, but the big takeaway isn't just the number of settings; it's that these reconstructed marginals reveal that layered entanglement structure we discussed, which is vital for those quantum networks.

Lev: From my side, I’m seeing how they use the inner product criterion to certify the dimensionality vector even when your experimental fidelity drops below zero point seven five, which is a key point for us running systems with real noise.

Kai: That means we can still verify the state's quality if our measurement setup isn't perfect, which is always going to be the case in a real lab setting.

Mira: It moves us beyond just checking if the state is entangled or not; it lets us quantify *how* entangled it is across all pairs, which helps reveal that layered entanglement structure we mentioned earlier.

Lev: That layered structure is what matters for building quantum networks, so this isn't just theoretical characterization; it’s a direct tool for designing better communication links.

The paper's summary: Mira: Now, looking at how they improve the original idea, they use this inner product criterion instead of just the simple fidelity-based criterion for certifying entanglement dimensionality.

Lev: That’s a big deal because using that inner product criterion shows a clear advantage over the fidelity-based criterion in detecting entanglement dimensionality vector even when experimental fidelity falls below zero point seven five.

Kai: So, if we look at what they built—that photonic four party entangled state with its specific dimensions of four by four by two by two—this method lets us use that marginal information to get more reliable results.

The paper's improvements: Kai: To wrap up on "Experimental High-Dimensional Quantum Overlapping Tomography," they showed that graph theory makes high dimensional tomography much more efficient and robust against noise when you look at systems like our four party state.

Lev: For real hardware, this means we can build better tools to verify the quality of our quantum connections without needing an impossibly high number of measurements.

Conclusion: Kai: So that was our rundown on "Experimental High-Dimensional Quantum Overlapping Tomography," where they use graph theory to get all six two body marginals from just twenty five measurement settings on a four party system.

Mira: Yeah, the main thing is that this method gives us a practical way to characterize the entanglement properties of all subsystems, which reveals that layered entanglement structure we discussed.

Lev: From an error correction standpoint, what I think is key is how they use that inner product criterion to certify the dimensionality vector even when your experimental fidelity drops below zero point seven five.

Kai: Right, and they also show how maximum likelihood estimation lets you reconstruct the actual physical density matrices from all those measurements.

Mira: So, it’s a whole pipeline where you use the graph theory for efficient measurement selection, then you use maximum likelihood estimation to figure out what the physical state actually is.

Lev: It means that if we want to build fault tolerant systems for these high dimensional states, we have a better way to verify the initial state before you start adding complex error correction codes.

Kai: It’s about moving from just getting a number to actually understanding the physical correlations present in the system when you run those experiments on your hardware.

Mira: That is important because it gives us a tangible tool for analyzing multipartite high dimensional entanglement, which is something we need for quantum networks.

Lev: I think this approach provides a more noise resilient route to certification than just relying on the simple fidelity-based criterion.

Kai: So, in summary, "Experimental High-Dimensional Quantum Overlapping Tomography" shows that graph theory makes high dimensional tomography much more efficient and robust against noise when you look at systems like our four party state.

Mira: It’s a practical method for quantifying entanglement properties across these higher dimensional systems that's useful for building better quantum networks.

Lev: For real hardware, this means we can build better tools to verify the quality of our quantum connections without needing an impossibly high number of measurements.

Kai: Next up, I’m looking at some work on how transport behaves along bosonic tight-binding chains connected to Markovian reservoirs, which is pretty different from this entanglement focus.

Haifei Wang, Rui Qu, *Zhengning Yang, Xiaodan Lyu, Haotao Zhu, Zitong Xu, Lianzhen Cao, Joel Yang, Otfried Günne† and Weibo Gao‡

Centre for Quantum Technologies, Singapore · Quantum Science and Engineering Centre (QSec) at Nanyang Technological University, Singapore · School of Electrical and Electronic Engineering at Nanyang Technological University, Singapore · Division of Physics and Applied Physics at the School of Physical and Mathematical Sciences at Nanyang Technological University, Singapore · School of Physics and Electronic Information at Weifang University · Engineering Product Development at Singapore University of Technology and Design · Naturwissenschaftlich-Technische Fakultütat Siegen, Universität Siegen

quant-ph

Submitted: 2026-08-16

Updated: 2026-11-18

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 97/100

The gist: Large-scale quantum systems have advanced rapidly via the exploration of more particles and higher dimensions, offering great potential for developing quantum technologies.

Key concepts

High-Dimensional Quantum Overlapping Tomography (QOT)
This is a technique used to reconstruct the reduced states (marginals) of complex, high-dimensional quantum systems. It leverages measurements taken in mutually unbiased bases (MUBs) across multiple parties to gather enough information to map out the system's entanglement structure, similar to how traditional tomography works but extended for higher dimensions.
Graph-Theoretic Formulation
The protocol models the quantum measurement process using graph theory. Specifically, it relates reconstructing all two-body marginals to covering the edges of a complete multipartite graph (KD⃗) with the minimum number of cliques. This mathematical mapping allows for an optimized, efficient selection of measurement settings.
Inner Product Criterion (IPC)
The IPC is a method used to certify entanglement dimensionality without relying solely on fidelity measurements. It evaluates the inner product between vectors derived from the marginals. The paper shows that this criterion can provide a stronger detection power for entanglement structure than traditional fidelity-based methods, even when experimental errors are present.
Mutually Unbiased Bases (MUBs)
MUBs are a set of orthonormal measurement bases where any two vectors from different bases have an overlap probability of exactly 1/d. By using MUB measurements, the protocol gains maximal information about the quantum state in a structured way, which is crucial for efficiently reconstructing the required marginals.

Terminology

Summary

Large-scale quantum systems have advanced rapidly via the exploration of more particles and higher dimensions, offering great potential for developing quantum technologies. p. 0 Here we propose high-dimensional quantum overlapping tomography based on a graph-theoretic formulation, which allows one to efficiently reconstruct few-body marginals of multipartite highdimensional quantum systems. p. 0 We experimentally realize it on a photonic four-party entangled state in a 4 × 4 × 2 × 2 system. p. 0 Using measurements in mutually unbiased bases, we reconstruct all six two-body marginals with only 25 projective measurement settings, compared with 94 and 225 settings for independent tomography of all two-body reduced states and full state tomography, respectively. p. 0 The reconstructed marginals reveal a layered entanglement structure vital for highdimensional quantum networks. p. 0

Protocol Formulation

The protocol exploits high-dimensional QOT based on graph theory for reconstructing all two-body marginals of the global quantum system. For a primepower dimension d, a complete set of MUBs contains (d + 1) orthonormal bases, with the property ⟨ψϕ⟩2 = 1/d for any two vectors ψ⟩ and ϕ⟩ from two different bases. p. 1 A global measurement setting is represented by choosing one vertex from each particle and connecting all of them, forming a complete subgraph, or a clique. p. 1 The optimal QOT protocol for reconstructing all two-body marginals is equivalent to covering all edges of KD⃗ with the smallest possible set of cliques. p. 1 For Kd+1,d+1,d+1, only (d+1)2 measurement settings are actually needed to reconstruct all two-body marginals. p. 1 The relevant graph is the complete multipartite graph KD⃗, where D⃗ = (d+1, …, d+1) with (d + 1) appearing n times. p. 1

Scalability Analysis

The scaling of high-dimensional QOT is shown to be logarithmic with the number of parties. The minimum number of settings required is denoted by Φ MUB k(n, d) = CAN(k, n, d + 1). p. 8 For fixed k, this gives Φ MUB k(n, d) = (d + 1) k O(log n). p. 8 This construction reduces the naive number of settings from (d+1) k n/k to (d + 1) k O(log n), changing the dependence on the number of parties n from polynomial to logarithmic. p. 9

Experimental Realization and Results

The authors experimentally prepare a photonic four-partite entangled state Λ⟩ = (1/2)(0011⟩ + 1100⟩ + 2211⟩ + 3300⟩). With only 25 mutually unbiased basis (MUB) settings, they reconstruct all six two-body marginals with an average fidelity of 0. p. 0876 ± 0.014 with respect to the target marginals. p. 0 The reconstructed marginals reveal the intended asymmetric pairwise entanglement structure useful for layered quantum communication in quantum networks. p. 1 Only ρAB shows significant non-zero off-diagonal terms, indicating a special correlation structure that may be useful for layered quantum communication where parties with higher security levels can share extra secret keys. p. 3

Noise Resilience and Certification

The marginals obtained from QOT enable a more noise-resilient route to certifying the entanglement dimensionality vector. Specifically, using marginal information, they can evaluate a lower bound on the Schmidt numbers via the inner-product criterion (IPC), which in theory has stronger detection power than the fidelity-based criterion (FBC) in most cases. p. 3 As shown in Fig. p. 1 S6, IPC can still certify the entanglement dimensionality vector when the FBC fails with experimental fidelity F < 0. p. 475.

Sample Complexity

The sample complexity for QOT is determined by Eq. (12), which shows N = O(σ 2 max / ε squared log K δ). p. 9 For fixed local dimension d and fixed marginal size k, the number of samples grows only logarithmically with the total number of parties. p. 9 This leads to a sample complexity of N = O(d 2k / ε squared k log n + log 1/δ). p. 9

Implementation Details

The implementation involves encoding the hybrid path–polarization states as Hu⟩ → 0⟩, Hl⟩ → 1⟩, Vu⟩ → 2», and Vl» →3». The MUB measurements are implemented by a measurement module shown in Fig. p. 2 S3. Maximum-likelihood estimation is used to reconstruct the physical density matrices from the experimental QOT data. p. 3

Conclusion

In summary, we reported the implementation of high-dimensional overlapping tomography optimized by a graph-theoretical formulation. We experimentally demonstrated it on a four-photon entangled quantum state with local dimensions of (4,4,2,2). p. 4 By reconstructing all two-body marginals with only 25 projective measurement settings, we quantified the entanglement properties of all two-body subsystems providing a practical characterization method for applications such as layered quantum communication networks. p. 4 Moreover, they showed that these marginals enable a noise-resilient approach to multipartite high-dimensional entanglement certification. p. 4

--- Page 1 ---

Experimental High-Dimensional Quantum Overlapping Tomography Haifei Wang, Rui Qu, Zhengning Yang, Xiaodan Lyu, Haotao Zhu, Zitong Xu, Lianzhen Cao, Joel Yang, Otfried G¨uhne and Weibo Gao <ref:2608.15795#pg12>

--- Page 2 ---

Quantum Science and Engineering Centre (QSec), Nanyang Technological University, Singapore 639798, Singapore <ref:2608.15795#pg13>

--- Page 3 ---

School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore <ref:2608.15795#pg14>

--- Page 4 ---

Division of Physics and Applied Physics, School of Physical and Mathematical Sciences, Nanyang Technological University, Singapore, 637371 <ref:2608.15795#pg15>

--- Page 5 ---

School of Physics and Electronic Information, Weifang University, Weifang 261061 <ref:2608.15795#pg16>

--- Page 6 ---

Engineering Product Development, Singapore University of Technology and Design, Singapore 487372 <ref:2608.15795#pg17>

--- Page 7 ---

Naturwissenschaftlich-Technische Fakult¨at, Universit¨at Siegen, Walter-Flex-Straße 3, 57068 Siegen, Germany <ref:2608.15795#pg18>

--- Page 8 ---

(Dated: August 18, 2026) <ref:2608.15795#pg19>

--- Page 9 ---

arXiv:2608.15795v1 [quant-ph] 16 Aug 2026

--- Page 10 ---

3 A covering array CA(N; k, n, v) is an N ×n array over an alphabet of size v such that every N ×k subarray contains all v k possible k-tuples at least once [1]

--- Page 11 ---

B. Scaling of measurement settings We now prove the scaling (d + 1)kO(log n) of the MUB-based construction

--- Page 12 ---

C. Confidence region and sample complexity In this part, we discuss how many copies of the state are required to reconstruct the desired marginals within a prescribed statistical accuracy

--- Page 13 ---

D. Maximum-likelihood estimation of the marginals We describe how the two-body marginals are reconstructed from the experimental QOT data

--- Page 14 ---

E. Noise resilience of the inner-product criterion for entanglement dimensionality certification Here we compare the inner-product criterion (IPC) the fidelity-based criterion (FBC) for certifying high-dimensional entanglement under added white noise

--- Page 15 ---

F. Data postprocessing for the experimental noisy states To investigate noise resilience, we recorded coincidence counts generated by independent white-noise sources coupled into the detectors

--- Page 16 ---

G. Maximum-likelihood estimation of the marginals Let X denote one of the six two-body subsystems, X ∈ S = OAB, AC, AD, BC, BD, CD

--- Page 17 ---

H. Confidence region and sample complexity We use the confidence-region method of Ref. p. 9 [6] to compare overlapping tomography schemes

--- Page 18 ---

I. Data postprocessing for the experimental noisy states The curves in Fig. p. 4 p. 4 4(b) in the main text are extracted from the depolarizing channel acting on the experimental state ρexp

--- Page 19 ---

J. Maximum-likelihood estimation of the marginals For a two-body subsystem X = X i, X j, we obtain the counts associated with local outcomes (oX i, oX j)

--- Page 20 ---

K. Confidence region and sample complexity The required measurement settings are obtained by solving the clique cover problem for K5,5,3,3

--- Page 21 ---

L. Implementation of mutually unbiased basis measurements A set of orthonormal bases is called mutually unbiased bases (MUB) if ⟨ψϕ⟩2 = 1/d holds for any two vectors ψ⟩ and ϕ⟩ belonging to two different bases

--- Page 22 ---

M. Data postprocessing for the experimental noisy states The fidelities, overlap ratios, and their uncertainties were evaluated from the combined counts using the same analysis pipeline as for the original data

--- Page 23 ---

N. Maximum-likelihood estimation of the marginals The MLE state is defined as the physical density matrix that maximizes this likelihood

--- Page 24 ---

O. Confidence region and sample complexity This procedure is applied to all six two-body subsystems

--- Page 25 ---

P. Implementation of mutually unbiased basis measurements We detail the implementation of the MUB measurements for our path-polarization qudit with local dimension d = 4

--- Page 26 ---

Q. Data postprocessing for the experimental noisy states The curves in Fig. 4(b) in the main text are extracted from the depolarizing channel acting on the experimental state ρexp

--- Page 27 ---

R. Maximum-likelihood estimation of the marginals For a two-body subsystem X = X i, X j, we obtain the counts associated with local outcomes (oX i, oX j)

--- Page 28 ---

S. Confidence region and sample complexity The required measurement settings are obtained by solving the clique cover problem for K5,5,3,3

--- Page 29 ---

T.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved AI systems can efficiently reconstruct few-body marginals of multipartite high-dimensional quantum systems using a graph theoretic formulation, allowing for reconstruction with only 25 projective measurement settings for a four-party system instead of 94 or 225 settings.

  2. The improved systems can provide noise resilient certification of multipartite high dimensional entanglement by utilizing inner product criterion which shows a clear advantage over the fidelity-based criterion in detecting entanglement dimensionality vector even when experimental fidelity falls below certain thresholds.

  3. The AI system can be used to quantify the entanglement properties of all two-body subsystems by reconstructing their corresponding physical density matrices using maximum likelihood estimation, enabling analysis of negativity to determine which subsystems are entangled with statistical significance.

Sources

Related papers