⚛️ quantum physics

Heisenberg-limited Bayesian phase estimation with low-depth digital quantum circuits

This paper proposes a scalable, low-depth digital quantum circuit scheme that achieves near-optimal Heisenberg-limited precision for Bayesian phase estimation by utilizing product GHZ states and adaptive (or optimized non-adaptive) measurements, while also addressing dynamic range extension and noise resilience.

Original authors: Su Direkci, Ran Finkelstein, Manuel Endres, Tuvia Gefen

Published 2026-02-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Su Direkci, Ran Finkelstein, Manuel Endres, Tuvia Gefen

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 are trying to tune a radio to a specific station, but the signal is fuzzy, and you don't know exactly which frequency you're looking for. You have a team of N tiny, super-sensitive helpers (qubits) to help you find the exact spot. The goal is to get the clearest signal possible with the least amount of effort.

This paper is about teaching that team of helpers how to find the "perfect station" (a phase) as accurately as physics allows, even when the starting conditions are messy.

Here is the breakdown of their breakthrough, using some everyday analogies:

1. The Problem: The "One-Size-Fits-All" Trap

In the world of quantum physics, there are two main ways to measure things:

The "Frequentist" Way: You know roughly where the station is. You just need to fine-tune it. The best tool for this is a GHZ state (think of it as a giant, perfectly synchronized choir of N voices singing in unison). This gives you super-precision, but it has a tiny range. If you are even slightly off-key, the choir gets confused and the signal breaks.
The "Bayesian" Way: You have no idea where the station is. It could be anywhere. This is like trying to find a needle in a haystack where the haystack is the size of a planet. The "choir" (GHZ state) is useless here because its range is too small.

The Dilemma: To get the best precision in the "Bayesian" (unknown) scenario, you theoretically need a very complex, messy state called a "sine state" and a super-complex measurement machine (like a Quantum Fourier Transform). But building these machines is incredibly hard, expensive, and often impossible with current technology.

2. The Solution: The "Lego Tower" Strategy

The authors asked: "Can we get that same super-precision using simple, easy-to-build tools?"

Their answer is yes. Instead of trying to build one giant, complex machine, they propose building a tower out of smaller Lego blocks.

The Blocks: Instead of one giant choir of N voices, they use many smaller choirs of different sizes (some with 2 voices, some with 4, some with 8, etc.).
The Strategy: They mathematically figure out exactly how many of each size block to use to create a "Lego tower" that mimics the behavior of the perfect, complex "sine state."
The Result: This "Lego tower" is much easier to build (it requires simple digital circuits) but performs almost as well as the impossible-to-build perfect machine.

3. The "Adaptive" Detective

Once they built their Lego tower, they needed a way to listen to it.

Old Way: Ask everyone to shout their answer at the same time and hope for the best. (This is "non-adaptive").
New Way (Adaptive): This is like a detective solving a mystery. You ask the biggest choir first. Based on their answer, you decide exactly what question to ask the next smaller choir. Then, based on that answer, you ask the next one.
Why it works: By adjusting your questions in real-time based on previous answers, you narrow down the possibilities incredibly fast. This allows them to reach the Heisenberg Limit—the absolute best precision the universe allows—using only simple tools.

4. The "Slow Motion" Trick (Fixing the "Phase Slip")

There is one big catch. If the "station" (the phase) is too far away, even the best detective gets confused because the numbers wrap around (like a clock hitting 12 and going back to 1). This is called a Phase Slip.

Imagine trying to measure a distance that is longer than your ruler. You have to mark the ruler, move it, and measure again.

The Old Fix: Use a bunch of "slow atoms" (helpers who move in slow motion) to measure the big picture, then use your fast helpers for the fine details. But this usually requires a lot of extra slow helpers, which is wasteful.
The New Fix: The authors invented a clever "rescaling" trick. They realized they could mathematically "shrink" the problem. They treat the slow atoms as if they were just part of the Lego tower, but with a different scale.
The Benefit: They can now measure huge distances (wide phase ranges) using a negligible number of extra "slow" atoms. It's like using a single, cleverly placed mirror to measure a room that is miles long, instead of needing a team of people to walk it out.

The Big Picture: Why Should We Care?

This isn't just about radio stations. This is about Atomic Clocks.

Atomic clocks are the most precise timekeepers in the universe, used for GPS, internet synchronization, and testing the laws of physics. However, they are currently limited by "noise" (jitter in the laser) and the difficulty of building complex quantum machines.

Before: To get better clocks, we needed impossible machines or huge numbers of atoms.
Now: This paper shows we can build better, more stable clocks using simple, low-depth circuits (like the ones in current quantum computers) and a smart "Lego" strategy.

In a nutshell: The authors found a way to get the "God-tier" performance of quantum physics using "Casual Friday" equipment. They replaced complex, fragile machinery with a smart combination of simple blocks and a step-by-step detective strategy, making ultra-precise timekeeping and sensing much more achievable for the real world.

1. Problem Statement

Quantum metrology aims to surpass the Standard Quantum Limit (SQL) of $1/\sqrt{N}$ and reach the Heisenberg Limit (HL) of $1/N$ in phase estimation. While optimal protocols exist, they face significant implementation barriers:

Complexity: Optimal initial states (e.g., "sine states" or spin-squeezed states) and optimal measurements (e.g., Quantum Fourier Transform, QFT) typically require deep quantum circuits (polynomial depth), which are currently infeasible for many platforms like neutral atom arrays or superconducting qubits.
Bayesian Constraints: In practical applications like atomic clocks, the phase $\phi$ is a stochastic variable with a prior distribution (often Gaussian) of non-zero width $\delta\phi$ . Optimal states for frequentist estimation (GHZ states) fail in the Bayesian regime when $\delta\phi > 1/N$ due to phase wrapping (aliasing) errors.
Scalability: Existing schemes using blocks of GHZ states often only work for specific, discrete numbers of qubits ( $N$ ) and assume uniform prior distributions, limiting their applicability to real-world atomic clocks.

The core question addressed is: Can fundamental precision limits (Heisenberg scaling) be saturated for arbitrary prior widths using only low-depth (logarithmic or constant) digital quantum circuits?

2. Methodology

The authors propose a two-step adaptive protocol that approximates optimal quantum interferometry (OQI) using simple, low-depth circuits:

A. State Preparation: Blocks of GHZ States

Instead of generating complex sine states, the authors approximate the optimal initial state using an ensemble of Greenberger-Horne-Zeilinger (GHZ) states with varying numbers of qubits.

Structure: The total $N$ qubits are partitioned into blocks, where each block $i$ contains $m_i$ copies of a GHZ state with $2^{k_i}$ qubits.
Optimization: For a given total qubit number $N$ and prior width $\delta\phi$ , the authors numerically optimize the partition $\{k_i, m_i\}$ to minimize the Bayesian Mean Squared Error (BMSE).
Key Insight: They demonstrate that these "blocks of GHZ states" can approximate the optimal initial state (sine state) almost perfectly if optimal measurements are available, regardless of the prior width.

B. Measurement Strategy: Local Adaptive Measurements

To avoid the high complexity of global QFT measurements, the authors employ local, adaptive measurements:

Process: The protocol measures GHZ blocks sequentially, starting from the largest block.
Adaptivity: Before measuring each subsequent block, a single-qubit rotation ( $\Phi_i$ ) is applied. The angle $\Phi_i$ is determined based on the outcomes of previous measurements to minimize the posterior variance.
Optimization: The rotation angles are optimized via gradient descent to minimize the BMSE analytically, assuming a Gaussian prior.

C. Addressing Phase Slip Errors (Phase Unwinding)

For wide prior distributions (common in atomic clocks where interrogation time exceeds laser coherence time), phase slip errors ( $\phi$ outside $[-\pi, \pi]$ ) degrade performance.

Slow Atoms: The authors introduce "slow atoms" that accumulate fractional phases ( $\phi/2, \phi/4, \dots$ ) to estimate the integer part of the phase ( $P$ ) and "unwind" the ambiguity.
Rescaling Transformation: They develop a mathematical rescaling technique that maps the problem of finding optimal partitions of slow atoms and GHZ states back to the standard GHZ partition problem. This allows the efficient allocation of atoms to extend the dynamic range with negligible overhead.

3. Key Contributions

Low-Depth Heisenberg Scaling: The scheme achieves Heisenberg scaling ( $1/N$ ) with a constant overhead (approx. 1.56) for any prior width, using only logarithmic-depth circuits for state preparation and local adaptive measurements.
Universal Applicability: Unlike previous schemes (e.g., Refs. [52, 53]) that only work for specific $N$ values, this protocol is optimized for any number of qubits $N$ and any Gaussian prior width $\delta\phi$ .
Efficient Phase Unwinding: The proposed method for extending dynamic range using slow atoms is significantly more efficient than existing non-adaptive or adaptive unwinding schemes, requiring fewer "slow" atoms to achieve the same sensitivity.
Experimental Roadmap: The paper outlines a feasible experimental realization using neutral atom arrays (e.g., Yb-171), leveraging mid-circuit measurements and feed-forward control.

4. Results

Performance vs. Prior Width: For $N=21$ qubits, the proposed scheme outperforms both Coherent Spin States (CSS) and fixed-block schemes across all prior widths. It achieves a metrological gain of $g \approx 2.29$ (3.59 dB) over CSS at the optimal prior width ( $\delta\phi \approx 0.7$ rad).
Scaling with N: As $N$ increases (up to 350), the scheme maintains Heisenberg scaling. The overhead relative to the theoretical OQI limit is calculated to be 1.56, which is lower than the 1.66 overhead of the varying-block scheme in Ref. [53].
Dynamic Range Extension: When phase slip errors are present (wide priors), the proposed phase unwinding protocol converges to the same sensitivity as the narrow-prior case with a negligible number of additional slow atoms. This enables atomic clock stability scaling as $\sigma_y(\tau) \propto \tau^{-1}$ even when the interrogation time exceeds the laser coherence time.
Noise Robustness: Simulations show the scheme remains superior to CSS even in the presence of amplitude damping and measurement errors, though the metrological gain decays exponentially with error rates. The authors suggest error detection (erasure conversion) can mitigate this.

5. Significance

This work bridges the gap between theoretical quantum metrology limits and near-term experimental capabilities.

Atomic Clocks: It provides a concrete pathway to building Heisenberg-limited optical atomic clocks using current digital quantum control technologies (tweezer arrays), potentially revolutionizing timekeeping and fundamental physics tests.
Quantum Algorithms: It establishes fundamental precision limits for quantum phase estimation algorithms implemented on low-depth digital circuits, showing that complex global operations (like QFT) are not strictly necessary to reach the Heisenberg limit if adaptive local strategies are used.
Scalability: By removing the constraint of specific qubit numbers and enabling arbitrary prior widths, the protocol is directly applicable to scalable quantum sensors and computing devices.

In summary, the authors demonstrate that by combining optimally partitioned GHZ states with adaptive local measurements, one can achieve near-optimal Bayesian phase estimation with low-depth circuits, effectively overcoming the limitations of current quantum platforms.