⚛️ quantum physics

Efficient direct quantum state tomography using fan-out couplings

This paper introduces and experimentally validates an efficient direct quantum state tomography scheme using fan-out couplings that achieves constant circuit depth and enables scalable state reconstruction and verification on superconducting quantum processors up to 20 qubits.

Original authors: Jaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim

Published 2026-04-07

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jaekwon Chang, Guedong Park, Hyunseok Jeong, Yong Siah Teo, Yosep Kim

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

Imagine you have a giant, invisible Rubik's cube made of pure light. This cube represents a quantum state. To understand how your quantum computer is working, you need to take a "picture" of this cube to see exactly how its pieces are arranged. This process is called Quantum State Tomography.

The Problem: The Impossible Photo Shoot

In the old way of doing this (Standard Tomography), taking a picture of a quantum system is like trying to photograph a spinning coin.

If you have just a few coins (qubits), you can take photos from every angle (top, side, front, diagonal) to figure out exactly where they are.
But as you add more coins, the number of angles you need to check explodes. For a system with 20 coins, you'd need to take millions of photos from different angles to get a complete picture. It's so expensive and slow that it becomes impossible for large systems.

The New Solution: The "Fan-Out" Flashlight

This paper introduces a clever new trick called Direct Quantum State Tomography (DQST) using something called a Fan-Out Coupling.

Think of the old method as trying to inspect a dark room by shining a flashlight on one corner at a time, then moving the light to the next corner, and so on. It takes forever.

The new method is like having a special flashlight (the "Meter Qubit") that can shine on all the corners of the room at once, but in a very specific, controlled way.

Here is how it works, step-by-step:

The Meter Qubit (The Flashlight): You prepare a single "helper" qubit (the meter) and put it in a superposition state (like a flashlight that is both on and off at the same time).
The Fan-Out (The Magic Beam): Instead of shining the light on one thing at a time, you use a "Fan-Out" gate. This is like a prism that splits your single flashlight beam so it hits all the system qubits simultaneously.
The Interaction: Because the beam hits everything at once, the helper qubit gets "entangled" with the whole system instantly.
The Readout: You measure the helper qubit. Because it touched everything at once, the result tells you specific details about the relationships between the qubits (the "density matrix elements") without needing to rebuild the whole picture from scratch.

Why is this a Game-Changer?

1. Constant Speed (The Elevator Analogy)
In the old method, if you wanted to check a bigger room, you had to add more stairs (circuit depth) to reach the corners. The bigger the room, the longer the climb.
In this new method, the "Fan-Out" gate is like an elevator that goes straight to the top floor regardless of how many floors there are. Whether you have 4 qubits or 20 qubits, the "circuit depth" (the time it takes to run the experiment) stays the same. It's constant!

2. The "Magic Mirror" Trick (Error Correction)
One of the biggest problems in quantum computing is noise (errors). Usually, if you run a circuit twice to check for errors, the noise gets worse.
However, this specific "Fan-Out" gate has a magical property: if you run it twice, it cancels itself out and becomes a "do nothing" operation (like looking in a mirror twice and seeing yourself exactly as you were).

Why this helps: Scientists can intentionally run the gate multiple times to "amplify" the noise, measure it, and then mathematically subtract it out. It's like turning up the static on a radio to figure out exactly what the static sounds like, so you can filter it out later.

The Real-World Test

The researchers tested this on a real quantum computer (IBM's superconducting processor):

Full Reconstruction: They successfully reconstructed the full state of a 4-qubit system using less than half the number of measurement settings required by the old method.
The Big Win (20 Qubits): They managed to verify a "GHZ state" (a highly entangled state of 20 qubits) using just one single circuit setup.
- Analogy: Imagine trying to verify if a choir of 20 singers is singing in perfect harmony. The old way required checking every singer individually from every angle. The new way let them listen to the whole choir with one microphone setup and instantly know if the harmony was real, even with 20 singers.

The Bottom Line

This paper presents a new way to "photograph" quantum systems that is:

Faster: It doesn't get slower as the system gets bigger.
Cheaper: It needs fewer measurement settings.
Smarter: It has a built-in feature that makes it easier to fix errors.

It's a major step forward in making quantum computers large enough to solve real-world problems, because we finally have a way to check if they are actually working correctly without getting lost in the math.

1. Problem Statement

Quantum State Tomography (QST) is essential for validating quantum devices but faces a critical scalability bottleneck. Conventional QST requires an informationally complete dataset, leading to an exponential increase in measurement overhead ( $3^n$ Pauli settings) and classical post-processing costs as the system size ( $n$ ) grows.
While "Direct Quantum State Tomography" (DQST) offers a route to estimate specific density-matrix elements without full reconstruction, previous implementations suffered from:

Scalability issues: Requiring experimentally demanding multi-controlled gates or multiple meter qubits.
Circuit depth: Circuit depth often scaling with system size, increasing susceptibility to noise.
Noise sensitivity: Weak-measurement approaches suffer from large statistical noise, while strong-measurement schemes often require complex settings.

2. Methodology

The authors propose a scalable DQST scheme that combines strong-measurement estimation with a fan-out coupling architecture.

Core Architecture:
- Single Meter Qubit: A single ancillary (meter) qubit is prepared in the $|+\rangle$ state.
- Fan-Out Coupling: The meter qubit interacts with an $n$ -qubit target system via a controlled- $U^k_{ES}$ gate. This gate acts as a "fan-out," where the meter controls parallel Pauli-X operations on a specific subset of system qubits defined by a binary vector $k$ .
- Commuting Interactions: Since all controlled-X operations commute, they can be executed in a single circuit layer. This ensures the circuit depth remains constant (independent of $n$ ), provided the hardware supports all-to-all or long-range connectivity (or via mid-circuit measurements in nearest-neighbor architectures).
Measurement Protocol:
- The system qubits are measured in the computational basis ( $Z$ -basis).
- The meter qubit is measured in the $X$ and $Y$ bases.
- Extraction: The expectation values of the meter ( $\langle X \rangle$ and $\langle Y \rangle$ ), conditioned on specific system outcomes, directly yield the real and imaginary parts of specific density-matrix elements $\langle a+k | \rho_s | a \rangle$ .
Full Reconstruction vs. Verification:
- Full Tomography: Requires $2^{n+1}-1$ distinct configurations (choices of $k$ ) to reconstruct the full density matrix. This is exponentially more efficient than standard QST ( $3^n$ ).
- Verification (GHZ States): For specific tasks like estimating the fidelity of an $n$ -qubit GHZ state, only one configuration ( $k = 11\dots1$ , i.e., $U = X^{\otimes n}$ ) is needed, regardless of system size.
Error Mitigation:
- The fan-out gate is involutory (applying it twice yields the identity). This property allows for straightforward Zero-Noise Extrapolation (ZNE) via digital gate folding without altering the target state preparation.
- Quantum Readout Error Mitigation (QREM) is applied to correct measurement errors.

3. Key Contributions

Constant-Depth Architecture: The proposal demonstrates that DQST can be implemented with constant circuit depth using a single meter qubit and fan-out interactions, decoupling complexity from system size.
Involutory Noise Scaling: The specific nature of the fan-out gate enables efficient noise scaling for error mitigation techniques like ZNE, a feature not easily available in standard multi-controlled gate approaches.
Unified Framework: The scheme unifies full state reconstruction and targeted verification (e.g., fidelity estimation) under a single theoretical and experimental framework.
Experimental Scalability: The authors successfully demonstrated the scheme on a superconducting quantum processor (IBM Quantum) for up to 20 qubits, a scale where standard tomography is infeasible.

4. Experimental Results

The team validated the scheme on the IBM Quantum Aachen processor:

Four-Qubit Full Reconstruction:
- Reconstructed density matrices for $|GHZ_4\rangle$ , $|0\rangle^{\otimes 4}$ , and $|+\rangle^{\otimes 4}$ .
- Efficiency: Achieved performance comparable to standard QST using only 31 circuits (DQST) versus 81 circuits (Standard QST).
- Fidelity: With readout error mitigation, fidelities reached $\approx 97.4\%$ (GHZ), $99.3\%$ ( $|0\rangle$ ), and $98.0\%$ ( $|+\rangle$ ), matching or exceeding standard QST results.
- Cross-Fidelity: The reconstructed matrices from DQST and standard QST showed cross-fidelities $>98\%$ , confirming consistency.
Scalable GHZ Fidelity Estimation (Up to 20 Qubits):
- Estimated the fidelity of GHZ states for $n = 4$ to $20$ using a single measurement configuration.
- Error Mitigation Impact: Without mitigation, fidelity dropped below the entanglement threshold ($0.5$) at $n=20$ . By applying both QREM and ZNE, the 20-qubit fidelity was restored to $53.2\%$ , successfully certifying Genuine Multipartite Entanglement (GME).
- Noise Scaling: The method successfully utilized the involutory property to scale noise (1-fold, 3-fold, 5-fold circuits) and extrapolate to the zero-noise limit.

5. Significance

Scalability: This work provides a practical pathway to characterizing large-scale quantum systems (20+ qubits) where traditional tomography is impossible due to resource constraints.
Hardware Agnostic: While demonstrated on superconducting qubits (nearest-neighbor connectivity), the scheme is theoretically applicable to trapped-ion and Rydberg-atom systems with all-to-all connectivity, where fan-out gates are native.
Verification Efficiency: It establishes a highly efficient protocol for verifying entanglement in large systems, crucial for benchmarking future quantum computers.
Error Mitigation Integration: The inherent compatibility of the fan-out gate with noise-scaling techniques offers a robust solution for extracting reliable data from noisy intermediate-scale quantum (NISQ) devices.

In conclusion, the paper presents a robust, experimentally verified method for direct quantum state tomography that significantly reduces measurement overhead and circuit depth, enabling the characterization and verification of quantum states at scales previously inaccessible to standard techniques.