Probing Quantum Information Scrambling via Local Randomized Measurements

This paper proposes a pragmatic paradigm for characterizing quantum information scrambling by deriving an analytical expression for the averaged accessible information (AAI) under local randomized measurements and demonstrating its ability to efficiently distinguish diverse dynamical behaviors, such as many-body localization and ballistic transport, using the classical shadow protocol.

The Gist

Imagine you have a giant, complex machine made of thousands of tiny, interconnected gears (this is your quantum system). You give one specific gear a little nudge. In a normal machine, that nudge might just jiggle the nearby gears. But in a quantum machine, that single nudge gets "scrambled." It spreads out so quickly and mixes with so many other gears that if you only look at the gear you nudged, or even a small cluster of nearby gears, the information about your original nudge seems to have vanished. It's hidden in the complex, entangled dance of the whole machine.

For a long time, scientists wanted to measure exactly how much of that original "nudge" information was still recoverable from a small part of the machine. The gold standard for this was a concept called Holevo information. Think of this as the "perfect detective" method. To find the maximum amount of hidden information, the detective would need to know exactly how the machine is moving and then choose the perfect, custom-made tool to measure it. The problem? In the real world, we can't build these perfect, custom tools. They are too hard to make and require knowing too much about the system beforehand.

The New Approach: The "Blind" Random Search

This paper proposes a smarter, more practical way to solve the mystery. Instead of trying to be a perfect detective with a custom tool, the authors suggest being a "blind" explorer with a bag of random tools.

They introduce a new metric called Averaged Accessible Information (AAI). Here is how it works:

1. Random Probes: Instead of one perfect measurement, you perform many measurements using random settings (like flipping a coin to decide which way to look at the gears).
2. Averaging: You take all the results from these random guesses and average them out.
3. The Result: Surprisingly, this "blind" average tells you almost exactly the same thing as the "perfect detective" method. It reveals how much information is still accessible in a small part of the system, even though you didn't know what you were looking for.

The Magic Trick: The "Shadow" Protocol

Measuring a quantum system usually requires taking a snapshot of the entire thing, which is incredibly slow and difficult. The authors use a clever trick called the Classical Shadow Protocol.

Imagine you want to know the shape of a giant, invisible statue. Instead of trying to photograph the whole thing at once, you shine a flashlight on it from many random angles and take quick, blurry snapshots of the shadows it casts. By combining these simple, random shadows, you can mathematically reconstruct the shape of the statue without ever seeing it directly.

In the paper, this means they can take a few random measurements on the whole system and use a computer to instantly calculate the "purity" (a measure of how mixed up the information is) of any small part they care about. This makes the process fast and efficient.

What They Found: Four Different "Dances"

The authors tested their new "blind probe" method on four different types of quantum systems to see how they scramble information. They found that their method could clearly distinguish between four very different behaviors:

1. The "Confined" Dance (Mixed-Field Ising Model): Imagine a ball tied to a string. If you push it, it moves a bit but is pulled back. In this system, the information spreads a little but gets trapped or "confined" by the system's rules. The authors' method saw this confinement clearly.
2. The "Bullet" Dance (Transverse-Field Ising Model): Imagine throwing a ball in a vacuum. It flies straight and fast. Here, the information travels ballistically (like a bullet) across the system without getting stuck. The method tracked this rapid spread perfectly.
3. The "Echo" Dance (PXP Model): Imagine a drum that, when hit, doesn't just fade away but keeps beating in a rhythmic pattern for a long time. This system has "quantum scars" that cause the information to revive and repeat itself. The authors' method caught these persistent echoes.
4. The "Frozen" Dance (Many-Body Localization): Imagine a room full of people who are so distracted by their own phones that they don't talk to anyone else. If you whisper a secret to one person, it never spreads. In this system, disorder freezes the information in place. The method showed that the information stayed stuck and never moved.

The Bottom Line

The paper claims that you don't need a "perfect" measurement to understand how quantum information scrambles. By using a "blind" approach—randomizing your measurements and averaging the results—you can get a highly accurate picture of what's happening. This bridges the gap between complex mathematical theories and what scientists can actually do in a real laboratory, allowing them to watch quantum information dance in real-time using simple, randomized tools.

Technical Summary

Technical Summary: Probing Quantum Information Scrambling via Local Randomized Measurements

Problem Statement
In isolated quantum many-body systems, local information typically scrambles irreversibly across the entire system, becoming hidden within highly entangled, non-local degrees of freedom. While the theoretical signature of this scrambling resides in global correlations, experimental characterization is inherently restricted to local probes. Existing metrics for diagnosing scrambling, such as Out-of-Time-Order Correlators (OTOCs) and operator size, have become experimentally accessible. However, from an information-theoretic perspective, the ultimate measure of scrambling is the Holevo information ($\chi$), which quantifies the maximum classical information extractable from a quantum ensemble. A critical limitation of the Holevo information is that its evaluation requires a state-dependent optimal measurement, which is generally unattainable in experimental settings. Conversely, theoretical lower bounds like subentropy ($Q$) rely on the full eigen-spectrum of the density matrix, necessitating full state tomography that scales poorly with system size. There remains a gap in identifying a sensitive, experimentally feasible probe that can decode complex quantum information dynamics without relying on optimal measurements or full state reconstruction.

Methodology
The authors propose a framework based on Averaged Accessible Information (AAI), denoted as $\chi_2$. This metric is derived by averaging the information gain over all possible measurement bases using the Haar-random measure.

1. Theoretical Derivation:

   • The authors define the AAI for an ensemble of mixed states generated by local perturbations.
   • By averaging the classical Rényi-2 mutual information over Haar-random measurement bases, they derive an analytical expression for the metric $Q_2(\rho)$, which depends solely on the purity of the reduced density matrix:
     $$Q_2(\rho) = \log\left(\frac{2}{1 + \text{Tr}(\rho^2)}\right)$$
   • The AAI is then defined as the difference in $Q_2$ between the average state and the individual states of the ensemble: $\chi_2(t) = Q_2(\sum p_i \rho_i) - \sum p_i Q_2(\rho_i)$.
   • The paper rigorously proves that this operationally derived $Q_2$ is strictly positive, additive, and Schur-concave (unlike the standard Rényi-2 entropy), and establishes its equivalence to a complex spectral function and a contour integral representation.

2. Experimental Protocol (Classical Shadows):

   • To efficiently extract $\chi_2$ without full tomography, the authors employ the classical shadows protocol using single-qubit randomized Pauli measurements.
   • This approach decouples the measurement phase from the target observables. A single set of randomized measurements on the entire system allows for the simultaneous prediction of purities for all possible local subsystems.
   • The authors utilize a lookup table approach to calculate the shadow kernel efficiently, bypassing complex matrix operations, and employ the "median-of-means" estimator to ensure robustness against statistical outliers.

Key Results
The framework was validated through numerical simulations across four paradigmatic quantum many-body models, demonstrating that $\chi_2$ can resolve distinct dynamical behaviors:

• Mixed-Field Ising Model (MFIM): The metric captures chaotic confinement, where a longitudinal field induces a linear confining potential that binds domain walls, leading to true many-body scrambling.
• Transverse-Field Ising Model (TFIM): In this integrable system, $\chi_2$ reveals ballistic transport of unconstrained domain-wall excitations, distinguishing it from the chaotic confinement of the MFIM despite similar initial decay rates.
• PXP Model: The metric successfully detects persistent oscillatory revivals characteristic of quantum many-body scars, indicating weak ergodicity breaking and non-thermal dynamics.
• Many-Body Localization (MBL): In the presence of strong disorder, $\chi_2$ shows strict localization and the absence of transport, reflecting strong ergodicity breaking. The dynamics are shown to be dominated by local integrals of motion ("l-bits"), with the perturbation remaining confined to a microscopic neighborhood for extended times.

The study further verifies that the AAI derived from finite random sampling of Clifford unitaries (a 2-design) converges to the theoretical Haar-averaged values, confirming the operational feasibility of the metric.

Significance and Claims
The paper claims to establish a "pragmatic paradigm shift" in characterizing quantum information dynamics. By replacing the requirement for optimal, state-dependent measurements with efficient, "blind" randomized local probes, the authors bridge the gap between abstract information-theoretic bounds (like the Holevo bound) and practical experimental implementations on quantum simulators.

The primary significance lies in demonstrating that the AAI ($\chi_2$), derived from subsystem purities accessible via classical shadows, serves as a high-fidelity proxy for the full Holevo information. It provides a robust, unbiased, and experimentally tractable method to distinguish between diverse non-equilibrium phenomena—including chaotic confinement, ballistic transport, scar revivals, and many-body localization—without the need for full state tomography or optimal measurements. This framework enables the systematic exploration of complex quantum information scrambling in programmable quantum simulators.