⚛️ quantum physics

Physical-Work Fluctuation Relations from Accessible Quantum Macrostates

This paper demonstrates that incorporating coarse thermodynamic information, such as endpoint mean energy and spatial records, into fluctuation relations creates a physically selected estimator that significantly reduces the sampling cost of calculating free-energy differences in nonequilibrium quantum systems compared to the standard Jarzynski equality.

Original authors: Borhan Ahmadi

Published 2026-10-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Borhan Ahmadi

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

Thermodynamics is the science of heat and energy, but it is also the science of how we describe things when we cannot see every single detail. Imagine trying to describe a bustling city. You could list the name, address, and mood of every single person, but that is impossible to manage. Instead, you might describe the average temperature, the total number of people, and how many are in the park versus the office. This is the power of thermodynamics: it lets us understand complex systems by focusing on a few big, accessible numbers rather than the chaotic microscopic details. For over a century, physicists have used this approach to predict how systems behave, even when they are being pushed out of balance by external forces.

One of the most powerful tools in this field is a rule that connects the work done on a system to its final state. If you push a system in a specific way, the energy you put in tells you something about the difference between where it started and where it ended up. However, there is a catch. To get an accurate answer using this rule, you often have to run the experiment thousands or millions of times. This is because the calculation relies on averaging, and the average is heavily influenced by very rare, unusual events that happen only once in a while. These rare events carry so much weight that if you miss even one of them, your final answer could be wildly wrong. Scientists have long searched for a way to get the same accurate answer without needing to run the experiment so many times.

A researcher has now found a way to do exactly that, using a method that does not change the physical experiment or the microscopic paths the system takes. Instead, they changed how they look at the data at the very end of the process. In their study, they focused on a small quantum system made of particles that can interact with one another. They pushed this system from a calm starting state to a chaotic, driven state, just as one would in a standard experiment. As the system moved, they tracked the energy changes for every single run. But here is the innovation: at the end of the run, before they calculated the final answer, they took a quick, coarse look at the system. They measured the average energy and a simple record of where the particles were located, without needing to know the exact position of every single particle.

This coarse measurement allowed them to define a "maximum-entropy state." In plain terms, this is the most likely state the system could be in, given only the limited information they chose to keep. It is like knowing the average temperature and the number of people in a room, and then assuming the most random, spread-out arrangement of people that fits those facts. The researcher used this reconstructed state as a reference point. They discovered that by comparing their standard calculation against this reference, they could cancel out the noise caused by those rare, difficult-to-find events. The result was a new way to estimate the energy difference that required far fewer experimental runs to reach the same level of confidence.

The researcher tested this idea on a specific model involving four particles moving on a short chain. In their simulations, the standard method required nearly six billion runs to get a reliable answer with a high degree of certainty. By using their new method, which incorporated the coarse endpoint information, they reduced that number to just under three billion runs. This is a massive saving in computational effort. Even more interestingly, they found that they did not need to match the reference state perfectly to get a benefit. If they allowed the reference to drift slightly away from the exact measured endpoint, the savings grew even larger. They found a sweet spot where a tiny loss in the precision of the reference state led to a much larger gain in statistical efficiency.

This work does not change the laws of physics or the definition of the work being done. The microscopic paths the particles took remain exactly the same, and the final energy difference they are trying to find is the same. The breakthrough lies entirely in the statistical analysis. The researcher showed that the information gathered at the end of the process acts as a powerful tool to filter out the noise. It is as if they found a way to weigh the results of a million coin flips by looking at the final shape of the pile, allowing them to ignore the rare, wild swings that usually throw off the average. By using the physical state of the system itself to guide the math, they turned a difficult sampling problem into a much more manageable one.

The study confirms that we can improve our understanding of non-equilibrium systems without needing to build better machines or run longer experiments. We simply need to be smarter about what information we keep and how we use it. The researcher demonstrated that even a small amount of accessible information about the final state can dramatically lower the cost of finding the answer. This suggests that in future experiments, whether in the lab or in complex simulations, scientists can achieve high precision with significantly fewer resources. The method provides a clear path forward for estimating energy differences in quantum systems, turning a task that was once computationally expensive into one that is far more efficient, all by listening more closely to the story the system tells at the very end.

Technical Summary: Physical-Work Fluctuation Relations from Accessible Quantum Macrostates

Problem Statement
The Jarzynski equality allows for the recovery of equilibrium free-energy differences ( $\Delta F$ ) from non-equilibrium work trajectories. However, the practical application of this relation is often hindered by severe sampling issues. The exponential average required by the equality is dominated by rare trajectories with large weights, necessitating an impractically large number of experimental runs to achieve a reliable estimate. While previous strategies have attempted to mitigate this by modifying driving protocols, introducing escorted dynamics, or optimizing Hamiltonian paths, this paper addresses the problem without altering the physical work protocol or the underlying microscopic trajectories. The central question is whether independently measured thermodynamic information at the non-equilibrium endpoint can serve as a statistical resource to improve the estimation of the standard Jarzynski free energy.

Methodology
The author proposes a framework that separates the microscopic description of the system from a coarse-grained thermodynamic description based on accessible observables.

Accessible Quantum Macrostates: For a finite isolated quantum system evolving from an initial canonical state $\rho_i$ to a non-equilibrium endpoint $\rho_f$ , the observer retains only the mean energy $E$ and the probabilities $p_r$ of a coarse projective record $R$ . These constraints define a maximum-entropy state $\bar{\rho}_{E,R}$ , which serves as a thermodynamic representative of the undisturbed endpoint.
Physical-Work-and-Record Fluctuation Relation: The author derives an exact fluctuation relation that incorporates both the standard Two-Point Measurement (TPM) work values ( $W_{mn} = E^f_m - E^i_n$ ) and the coarse record $r$ . This relation defines a trajectory variable $\sigma_{\text{phys},R}$ whose exponential average is exactly unity. This variable includes terms for the work, the change in the maximum-entropy multipliers, and corrections for the non-commutativity of the final energy measurement and the record measurement.
Control Variate Strategy: The derived fluctuation relation belongs to a family of exact relations parameterized by $\vartheta = (\beta, \alpha)$ . The specific member matching the physical endpoint ( $\vartheta_{\text{th}}$ ) is selected by physical measurement rather than statistical optimization. This member, having a known mean of zero for the variable $C_\vartheta = e^{-\sigma_\vartheta} - 1$ , is used as a control variate for the standard Jarzynski estimator. By subtracting a scaled version of this control from the Jarzynski estimator, the variance of the free-energy estimate is reduced without discarding trajectories or changing the target $\Delta F$ .
Optimization of Fidelity vs. Sampling: The author investigates whether deviating slightly from the exact physical endpoint (allowing a relative entropy distance $\delta$ ) can further reduce sampling costs. They optimize the control parameters and the scaling coefficient $c$ to minimize the finite-confidence sampling cost.

Key Results
The study was validated using a finite Bose-Hubbard system (4 bosons on a 4-site open chain) undergoing a driven work stroke.

Decomposition of Fluctuations: The mean of the new trajectory variable $\langle \sigma_{\text{phys},R} \rangle$ was shown to decompose into three distinct terms: the relative entropy between the true state and the thermodynamic representative (unresolved information), a term arising from the measurement backaction on the record ( $\Delta_{\text{meas}}$ ), and a non-commutativity correction ( $Q_{nc}$ ).
Sampling Reduction: Using the endpoint-matching thermodynamic member as a control variate reduced the required number of production runs for a fixed confidence level from $5.6463 \times 10^9$ (direct Jarzynski) to $3.5873 \times 10^9$ .
Square-Root Trade-off: Allowing a small controlled departure from the exact thermodynamic endpoint ( $\delta > 0$ ) yielded further gains. At $\delta = 0.01$ , the optimized control reduced the sampling cost by an additional 17.78% relative to the exact match, achieving a total reduction of 47.76% compared to direct sampling.
Information-Sampling Frontier: The relationship between the deviation $\delta$ and the reduction in sampling cost follows a square-root law ( $1 - N^*/N_{\text{th}} \propto \sqrt{\delta}$ ) near the physical endpoint. This indicates that a small loss in thermodynamic fidelity yields a disproportionately large statistical gain.
Microscopic Mechanism: The variance reduction is attributed to the control variate's correlation with rare, high-energy initial sectors that dominate the variance of the standard Jarzynski estimator. The control effectively subtracts fluctuations from these rare events without discarding the trajectories.

Significance and Claims
The paper claims to establish a new strategy for improving free-energy estimation that relies on "accessible quantum macrostates" rather than modifying the physical dynamics.

Exact Statistical Control: The primary contribution is demonstrating that a thermodynamic state, independently fixed by a few measurements of the actual non-equilibrium endpoint, can serve as an exact control variate for the ordinary Jarzynski estimator. This provides exact statistical control even when the final energy and retained record do not commute.
Physical Selection: Unlike previous methods that optimize parameters purely for statistical convergence, the proposed control is "physically selected" by the measured endpoint, ensuring the method remains grounded in the system's actual thermodynamic state.
Efficiency without Alteration: The method substantially lowers the finite-confidence sampling cost in a finite interacting quantum system without changing the work protocol, the microscopic trajectories, or the definition of work.
Trade-off Principle: The work reveals a specific "information-sampling frontier" where a square-root onset allows for a controlled trade-off between thermodynamic fidelity and statistical efficiency, suggesting that minimal loss of endpoint matching can produce significant statistical gains.

The author emphasizes that this approach keeps the microscopic and thermodynamic descriptions separate, utilizing the coarse-grained information only to construct a statistical tool for the fine-grained work data. The results are presented as model-conditioned benchmarks, with the understanding that experimental implementation would require accounting for the cost of the initial calibration and pilot runs used to fix the control parameters.

Technical Summary: Physical-Work Fluctuation Relations from Accessible Quantum Macrostates

More like this