⚛️ quantum physics

Information conservation relations for weak measurement and its reversal

This paper derives exact, outcome-resolved information conservation relations for multilevel, decaying quantum systems under continuous weak monitoring and its reversal, establishing quantitative trade-offs that unify the understanding of information flow in open quantum dynamics.

Original authors: Yusef Maleki, Luis D. Zambrano Palma, M. Suhail Zubairy

Published 2026-01-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Yusef Maleki, Luis D. Zambrano Palma, M. Suhail Zubairy

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 guess what's inside a sealed, dark box. You can't open it, but you have a sensitive microphone outside that listens for a specific sound: a "click" that happens if a particle escapes the box.

This paper is about a very specific game of "20 Questions" played with quantum particles, where the rules of the game reveal a hidden balance sheet of information. Here is the breakdown of their findings using everyday analogies.

The Setup: The Leaky Box

Imagine a quantum system (like a tiny atom) as a leaky box.

The Leak: The box has a small hole. If the particle inside is "excited" (energetic), it might leak out. If it leaks, your detector outside hears a click.
The Silence: If the particle stays inside, the detector hears nothing (a "null result" or "no-click").
The Catch: In the quantum world, even silence tells you something. If you wait a long time and hear no click, you become more confident that the particle is likely in a "safe" state (the ground state) rather than an "energetic" state that would have leaked out.

The Core Discovery: The Information Ledger

The authors discovered that information isn't created or destroyed; it just gets shuffled around. They found a mathematical "conservation law" for this information, similar to how money in a bank account must balance (Deposits = Withdrawals + Remaining Balance).

They looked at three main scenarios:

1. The "Silence" Scenario (No Clicks)

When the detector stays silent, you gain information. But where does this information come from?

The Analogy: Imagine you are betting on a horse race. If a horse that is known to be fast (the excited state) doesn't run, you gain information that the slower horse (the ground state) is likely the winner.
The Balance: The paper shows that the information you gain from the silence is split into two parts:
1. The Update: How much your belief about the specific state of the system changed.
2. The Decay Cost: The "cost" of the time passing without a leak.
The Rule: The total information from the silence = (How much your guess changed) + (The information lost to the passage of time/decay). It's a perfect ledger; nothing is missing.

2. The "Undo" Scenario (Reversal)

What if, after the silence, you try to "rewind" the system to its original state?

The Analogy: Imagine you are trying to un-bake a cake. Sometimes you can reverse the process, but the chance of success depends on how much time has passed and how much the "leak" has happened.
The Balance: The authors found that the information gained from the silence is also linked to the probability of successfully reversing the process.
The Rule: If you know how likely you are to successfully "undo" the measurement, you can calculate exactly how the information is distributed. The "cost" of trying to reverse the system is directly tied to the information you gained from the silence. It's like a trade-off: the more information you gain from the silence, the harder it is to reverse the process perfectly.

3. The "Click" Scenario (Multiple Clicks)

What if the detector does click? What if it clicks once, twice, or three times?

The Analogy: Imagine the box is now a multi-story building. If you hear one click, you know a particle fell from a high floor. If you hear three clicks, you know it fell from the very top.
The Balance: The paper extends their rule to these "click" events. They found that even when the detector registers a specific number of clicks, the information balance still holds.
The Twist: When you get a click, the information comes from a mix of sources:
- The fact that a click happened.
- The fact that the particle didn't leak further (no decay).
- A "confusion tax": If you hear one click, you might not know exactly which floor it came from (was it floor 2 or floor 3?). This uncertainty reduces the total information you can extract. The paper quantifies this "confusion tax" and adds it to the balance sheet.

The Big Picture

The authors didn't just look at simple two-state systems (like a coin flip). They looked at complex systems with many levels (like a ladder with many rungs).

Their main conclusion is that information flow in these open quantum systems is predictable and conserved.

Whether the detector is silent or clicking.
Whether you try to reverse the process or not.
Whether the system has 2 levels or 100 levels.

There is always a strict equation that balances the information you gain, the information lost to the environment (decay), and the information required to reverse the process. It's like a universal accounting rule for the quantum world: Information is never lost; it is just moved between the system, the detector, and the environment.

What This Means (According to the Paper)

The paper claims this provides a "unified account" of how information moves in these systems. It helps scientists understand exactly how much they know about a system at any given moment and how much "effort" (in terms of probability) it would take to go back to the start. It doesn't promise to build a new computer or cure a disease, but it provides the fundamental "rulebook" for how information behaves when we peek at quantum systems without breaking them.

Technical Summary: Information Conservation Relations for Weak Measurement and Its Reversal

Problem Statement
The paper addresses the lack of a complete, quantitative account of information flow in weak measurement protocols, particularly within multilevel, decaying quantum systems under continuous monitoring. While the trade-off between information gain and state disturbance is well-established in projective measurements, and the role of null outcomes (no-click events) in weak measurements is recognized, a unified conservation framework describing how information is distributed among the system, the measurement record, and the decay channel remains underdeveloped. Specifically, the authors seek to derive exact relations that hold for individual measurement outcomes, extending beyond average behaviors to include the probabilistic reversal of weak measurements and finite-count outcomes (arbitrary photon numbers).

Methodology
The authors employ an information-theoretic framework applied to open quantum dynamics. The methodology proceeds through several stages of generalization:

Single-Qubit Model: The analysis begins with a qubit modeled as a superposition of vacuum and one-photon Fock states in a cavity. The system is monitored by a detector outside the cavity. The authors utilize the photon-detection measurement operator $M_k$ to describe the evolution. They focus on the "null-result" (no-click) case ( $k=0$ ), deriving the posterior state and calculating pointwise information gains defined as the difference between prior and conditional information ( $\Delta I = I_{prior} - I_{posterior}$ ).
Extension to Multilevel Systems: The formalism is generalized to an $N$ -level system in the Fock basis. The authors derive expressions for the probability of null detection and the resulting conditional probabilities for each energy level $n$ .
Average Information and Relative Entropy: The authors calculate the average information gain by averaging over the posterior distribution. They identify this average gain as the relative entropy (Kullback-Leibler divergence) between the initial and final state distributions.
Reversal Protocol: Incorporating the concept of weak measurement reversal (following the approach of Sun et al.), the authors define a reversal success probability $p(\text{rev})$ . They express this probability in information-theoretic terms to establish a balance relation involving the cost of reversal.
Finite-Count Outcomes: Finally, the framework is extended to arbitrary $k$ -click events (where the detector registers $k$ photons). This involves generalizing the measurement operators and deriving a conservation relation that accounts for decay, no-decay, and the indistinguishability of photon-number states.

Key Contributions and Results

Exact Conservation Relations for Null Outcomes: For a multilevel system under continuous monitoring with no-click detection, the authors derive an exact conservation-type relation:
$I(y_0) = \Delta I(x_n|y_0) + n I(\text{decay})$
Here, $I(y_0)$ is the information content of the no-click event, $\Delta I(x_n|y_0)$ is the pointwise information gain regarding the system being in state $|n\rangle$ , and $n I(\text{decay})$ represents the information associated with the decay channel. This equation demonstrates that the information carried by the null outcome is distributed between the gain in knowledge about the system's state and the information lost to the decay process.
Average Information Balance: By averaging over the posterior distribution, the authors establish a compact relation linking the total information of the null result to the relative entropy and the average photon number:
$I(y_0) = D(p(x_n|y_0) \parallel p(x_n)) + \langle n \rangle I(\text{decay})$
This confirms that the information gain from a dark count is bounded by the relative entropy, quantifying the distinguishability between the prior and posterior states.
Reversal and Information Trade-offs: When an explicit reversal is attempted, the authors derive a relation incorporating the reversal success probability:
$I(y_0) = N I(\text{decay}) - I(\text{rev})$
Combining this with the pointwise gain, they show that the information content of the null result can be reconstructed from the information gain and the "cost" of reversal. Notably, they find that the informational structure of higher energy levels is fully determined by the contributions of the ground state and the first excited state.
Generalized Relations for $k$ -Click Events: For outcomes where $k$ photons are detected, the authors derive a generalized conservation relation:
$I(y_k) = D(p(x_n|y_k) \parallel p(x_n)) + (\langle n \rangle - k) I(\text{decay}) + k I(\text{no decay}) - \langle I(W) \rangle$
This relation accounts for the information from the decay process, the no-decay process, the specific detection of $k$ clicks, and a correction term $\langle I(W) \rangle$ representing the information cost due to the indistinguishability of underlying photon-number states that could produce the same $k$ -click outcome.

Significance
The paper claims to provide a unified, outcome-resolved account of information flow in monitored open quantum dynamics. By deriving exact, conservation-type relations that hold for each specific outcome (whether null or finite-count), the work offers a deeper understanding of how information is partitioned between the system, the measurement record, and the environment.

The authors posit that these results sharpen the principles for analyzing and controlling weak measurements and their reversals in realistic, noisy settings. The framework explicitly links the change in a system's information content to the entities to which that information is distributed (the decay channel, the reversal probability, or the detector). While the paper does not propose new experimental setups or specific applications beyond the theoretical analysis, it establishes a rigorous quantitative foundation for future studies on open-system dynamics, potentially aiding in the characterization and mitigation of coherence loss in quantum technologies. The authors suggest that this framework invites extensions to non-Markovian baths, continuous-variable systems, and adaptive feedback strategies, though these are presented as potential directions rather than immediate results of the current work.