Efficient direct quantum state tomography using fan-out couplings

Jaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim · 2026

Read Efficient direct quantum state tomography using fan-out couplings by Jaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-04454.txt.

Introduction

Chang and colleagues introduce a direct quantum state tomography scheme mixing strong-measurement estimation with a fan-out coupling architecture. Their scheme lets system qubits interact with a single meter qubit through commuting links, reaching constant circuit depth. The fan-out coupling reduces to identity on repeat, enabling simple noise scaling for quantum error mitigation. Tests on a quantum chip via the IBM Quantum Platform rebuild a four-qubit state. They estimate single-circuit GHZ-state fidelity up to 20 qubits with error control, matching standard tomography efficiently.

Catalog record

CREATORJaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE57641
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.04454

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Efficient direct quantum state tomography using fan-out couplings Jaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim Abstract Characterizing quantum states is essential for validating quantum devices, yet conventional quantum state tomography becomes prohibitively expensive as system size grows. Direct tomography offers a distinct route by enabling selective access to individual complex density-matrix elements, with a particular advantage for sparse target states and some verification tasks. Here we introduce a direct quantum state tomography scheme combining strong-measurement estimation with a fan-out coupling architecture. It enables mutually commuting interactions between system qubits and a single meter qubit, thereby achieving constant circuit depth, independent of system size. Notably, the involutory fan-out coupling reduces to the identity under repetition, enabling straightforward noise scaling for quantum error mitigation. We experimentally validate the scheme on a superconducting quantum processor via the IBM Quantum Platform, demonstrating four-qubit state reconstruction and single-circuit GHZ-state fidelity estimation up to 20 qubits with error mitigation. Consistent results with standard tomography and improved efficiency establish our scheme as a promising approach to reconstructing full quantum states and scalable verification tasks. Introduction Quantum state tomography lies at the core of quantum characterization and verification by providing complete state information 18 . However, faithful reconstruction requires informationally complete data, resulting in exponential measurement overhead and substantial classical post-processing costs with increasing system size 1 . To mitigate these overheads, a priori information about the target, such as sparsity 45 ; 34 or low rank 11 ; 9 ; 6 , is often leveraged. In addition, adaptive measurement 23 ; 39 ; 32 and machine-learning-assisted approaches have emerged as viable directions 53 ; 46 ; 51 ; 5 . For more limited objectives, sampling-efficient verification protocols have been developed for some state classes 2 , including direct fidelity estimation 8 ; 43 and shadow tomography 21 ; 22 ; 44 . Direct quantum state tomography (DQST) provides a unified framework for both full state reconstruction and verification tasks. It enables the estimation of individual complex density-matrix elements without requiring full state reconstruction. This element-selective strategy is particularly advantageous for sparse quantum states and well suited to verification tasks, including fidelity estimation 29 , entanglement witness 12 ; 14 , and coherence measure 49 . Moreover, on platforms where switching measurement settings is more costly than increasing measurement shots 36 , DQST can outperform randomized-measurement-based verification protocols by targeting specific matrix elements with minimal settings. Early direct tomography was introduced through the weak-value framework based on sequential measurements 37 ; 33 ; 30 ; 40 ; 59 . When the first measurement is sufficiently weak, the disturbance to the system remains minimal, such that otherwise incompatible sequential measurements can still provide meaningful information 7 ; 29 . For example, a system in |\psi\rangle is weakly coupled to a meter measuring the position observable |x\rangle\langle x| and subsequently post-selected in a momentum state |p\rangle . This procedure produces a meter shift proportional to the complex amplitude of the system state \langle p|x\rangle\langle x|\psi\rangle 37 . Although originally formulated for pure states, the framework was extended to directly extract density-matrix elements 38 ; 52 . Nevertheless, the weak coupling transfers only limited information and suffers from large statistical noise, which later motivated strong-measurement schemes at the cost of additional settings 54 ; 58 ; 4 ; 42 ; 60 . While the scalability of DQST has been demonstrated using system-specific high-dimensional interactions 40 ; 58 ; 59 , extending it to general circuit-based implementations typically requires either experimentally demanding multi-controlled gates 60 ; 31 ; 41 or multiple meter qubits 4 ; 42 . In this work, we propose an experimentally scalable DQST scheme in which a single meter qubit is strongly coupled to multiple system qubits via a single fan-out gate 24 . This architecture allows the circuit depth to be compressed to a constant 35 ; 15 ; 3 ; 48 ; 17 , independent of system size, while providing programmable access to arbitrary subsets of the density-matrix elements. In addition, the involutory fan-out gate reduces to the identity under repetition, enabling straightforward noise scaling for quantum error mitigation 50 ; 28 ; 10 ; 19 . To benchmark its performance, we experimentally demonstrate our DQST scheme via full state reconstruction of four-qubit states on a superconducting quantum processor using the IBM Quantum Platform 25 . In addition, to demonstrate efficient verification, we estimate GHZ-state fidelity for up to 20 qubits using a single circuit with quantum error mitigation. Consistent results with standard tomography and improved efficiency establish our scheme as a promising approach to reconstructing full quantum states and scalable verification tasks. Results Schematic of DQST. Figure 1 a illustrates a quantum circuit implementing our DQST scheme. A meter qubit is first prepared in |+\rangle_{\mathrm{m}}=(|0\rangle_{\mathrm{m}}+|1\rangle_{\mathrm{m}})/\sqrt{2} using a Hadamard gate. It then interacts with an n -qubit target system \rho_{\mathrm{s}} via a controlled- U^{\mathbf{k}}_{\mathrm{ES}} gate: \displaystyle\lambda^{\mathbf{k}}_{\mathrm{sm}} \displaystyle=\frac{1}{2}\Bigl(\rho_{\mathrm{s}}\otimes|0\rangle\langle 0|_{\mathrm{m}}+(U^{\mathbf{k}}_{\mathrm{ES}}\rho_{\mathrm{s}})\otimes|1\rangle\langle 0|_{\mathrm{m}} (1) \displaystyle\quad\;+(\rho_{\mathrm{s}}U_{\mathrm{ES}}^{\mathbf{k}\,\dagger})\otimes|0\rangle\langle 1|_{\mathrm{m}}+(U^{\mathbf{k}}_{\mathrm{ES}}\rho_{\mathrm{s}}U_{\mathrm{ES}}^{\mathbf{k}\,\dagger})\otimes|1\rangle\langle 1|_{\mathrm{m}}\Bigr). The system-meter output state \lambda^{\mathbf{k}}_{\mathrm{sm}} shows that measuring the meter qubit in the Pauli- X or Y basis enables a coherent superposition of left- and right-actions of U^{\mathbf{k}}_{\mathrm{ES}} on the system density matrix: \displaystyle\langle X_{\mathbf{a}}^{\mathbf{k}}\rangle=\mathrm{Tr}\!\left[\lambda^{\mathbf{k}}_{\mathrm{sm}}\,|\mathbf{a}\rangle\langle\mathbf{a}|_{\mathrm{s}}\otimes X_{\mathrm{m}}\right]=\tfrac{1}{2}\langle\mathbf{a}|\rho_{\mathrm{s}}U_{\mathrm{ES}}^{\mathbf{k}\,\dagger}+U_{\mathrm{ES}}^{\mathbf{k}}\rho_{\mathrm{s}}|\mathbf{a}\rangle, \displaystyle\langle Y_{\mathbf{a}}^{\mathbf{k}}\rangle=\mathrm{Tr}\!\left[\lambda^{\mathbf{k}}_{\mathrm{sm}}\,|\mathbf{a}\rangle\langle\mathbf{a}|_{\mathrm{s}}\otimes Y_{\mathrm{m}}\right]=\tfrac{i}{2}\langle\mathbf{a}|\rho_{\mathrm{s}}U_{\mathrm{ES}}^{\mathbf{k}\,\dagger}-U_{\mathrm{ES}}^{\mathbf{k}}\rho_{\mathrm{s}}|\mathbf{a}\rangle, (2) where \mathbf{a}\in\{0,1\}^{n} denotes the n -bit measurement outcome in the computational basis. To extract matrix elements of the system state, we employ a controlled- U_{\mathrm{ES}}^{\mathbf{k}} gate such that \langle\mathbf{a}|U_{\mathrm{ES}}^{\mathbf{k}}=\langle\mathbf{a}+\mathbf{k}| (mod 2), where \mathbf{k}\in\{0,1\}^{n} specifies the system qubits on which Pauli- X acts. With this choice, the real and imaginary matrix elements are obtained from meter expectation values conditioned on projection onto |\mathbf{a}\rangle : \displaystyle\langle X_{\mathbf{a}}^{\mathbf{k}}\rangle=\mathrm{Re}\bigl[\langle\mathbf{a}+\mathbf{k}|\rho_{\mathrm{s}}|\mathbf{a}\rangle\bigr]=\mathrm{Re}\bigl[\langle\mathbf{a}|\rho_{\mathrm{s}}|\mathbf{a}+\mathbf{k}\rangle\bigr], \displaystyle\langle Y_{\mathbf{a}}^{\mathbf{k}}\rangle=\mathrm{Im}\bigl[\langle\mathbf{a}+\mathbf{k}|\rho_{\mathrm{s}}|\mathbf{a}\rangle\bigr]=-\mathrm{Im}\bigl[\langle\mathbf{a}|\rho_{\mathrm{s}}|\mathbf{a}+\mathbf{k}\rangle\bigr]. (3) Sample complexity and additional details are provided in the Methods and Supplementary Note 1. The controlled- U_{\mathrm{ES}}^{\mathbf{k}} operation can be implemented as a fan-out gate, where a single meter qubit acts as a control and conditionally flips multiple system qubits via parallel CNOT operations. For example, the set of matrix elements \langle\mathbf{a}|\rho_{\mathrm{s}}|\mathbf{a}+101\rangle is obtained using U_{\mathrm{ES}}^{101}=X_{1}I_{2}X_{3} , implemented via CNOT gates between the meter qubit and system qubits 1 and 3 (see Fig. 1 b ). Since all control-target interactions commute, they can be executed within a single circuit layer, rendering the circuit depth, in principle, independent of both the system size and the specific choice of U^{\mathbf{k}}_{\mathrm{ES}} . A single-depth fan-out gate can be experimentally implemented by simultaneously activating interactions in ion-trap and Rydberg-atom systems with all-to-all or long-range connectivity 35 ; 15 . Even in superconducting qubit systems with nearest-neighbor connectivity, it can be realized at constant depth via mid-circuit measurements 3 ; 48 ; 17 . Figure 1 : Schematic of DQST. a , Circuit diagram for matrix-element estimation of an n -qubit system \rho_{\mathrm{s}} . The meter qubit is prepared in |+\rangle_{\mathrm{m}} and coupled to the system via a controlled- U_{\mathrm{ES}}^{\mathbf{k}} gate. The system qubits are measured in the computational basis, while the meter qubit is measured in the X and Y bases to access the real and imaginary parts, respectively (see Eq. ( 3 )). b , Example of a controlled- U_{\mathrm{ES}}^{\mathbf{k}} gate with \mathbf{k}=101 . c , Accessible matrix-element subsets for each U_{\mathrm{ES}}^{\mathbf{k}} . The subset corresponding to b is highlighted in yellow. Figure 2 : Density matrix reconstruction results. Four-qubit density matrices of the |\mathrm{GHZ}_{4}\rangle

Read the complete text