Boltzmann-like occupation of nonequilibrium steady states on dense disordered networks
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: We're looking at the paper: "Boltzmann-like occupation of nonequilibrium steady states on dense disordered networks".
Mira: Here's the gist of the paper. Boltzmann-like occupation of nonequilibrium steady states on dense networks proves that NESS on large dense networks have Boltzmann-like occupation despite extensive entropy production.
Kai: This finding is significant because it extends the Boltzmann distribution to nonequilibrium steady states (NESS) and establishes that the active-matter heuristic of “low rattling” is asymptotically exact for such NESS.
Mira: Modeling Nonequilibrium Steady States The study models the steady state as a Markov jump process on discrete states where the transition rate from state j to state i is given by W ij = exp(E j X ij).
Kai: The authors define the column sums of W = (W ij) i≠j as-W jj = w j = sum i≠j W ij.
Paper summary: Kai: So we're looking at this paper today, "Boltzmann-like occupation of nonequilibrium steady states on dense disordered networks." Basically, they’re trying to take the standard Boltzmann distribution and apply it to systems where things aren't in thermal equilibrium—nonequilibrium steady states, or NESS.
Mira: That’s the big problem in statistical physics right there. The paper claims that even on these large dense networks, we still see Boltzmann-like occupation, even though there’s a lot of entropy production happening.
Lev: Because the authors show that this distribution holds when the state space is large enough, specifically when you look at the maximum absolute relative error between the NESS and equilibrium distributions going to zero as N goes to infinity <ref:2606.14542#pg2>.
Kai: It’s not just that it converges; they show that this convergence happens in terms of all sorts of things, like the relative entropy and total variation distance also going to zero <ref:2606.14542#pg2>. That tells us it’s a very robust result as the system gets bigger.
Mira: And on top of that, they show that the active-matter heuristic about "low rattling" is actually asymptotically exact for these NESS <ref:2606.14542#pg3>. Which means the states you find themselves in are those you leave more slowly, and this explanation extends to a broader class of states called equiaccessible steady states <ref:2606.14542#pg1>.
Lev: From an engineering standpoint, that suggests that we can use this concept of low rattling to predict how occupation will look without needing the full complexity of the nonequilibrium dynamics <ref:2606.14542#pg3>. It simplifies things significantly for running simulations or modeling real hardware <ref:2606.14542#pg3>.
Kai: The paper also introduces a nonequilibrium fluctuation-dissipation theorem, which is interesting because it implies that just measuring how the occupation responds to energy changes isn't enough to tell if the steady state is actually thermal equilibrium <ref:2606.14542#pg3>.
Mira: They show that for any energy perturbation you make, the change in occupation scales like one plus a term that goes to zero as N gets larger, which is dE i pi versus dE i pi eq <ref:2606.14542#pg3>. This links the static response to things that are not necessarily in equilibrium.
Lev: What this means for hardware is that we can't rely solely on a static measurement of how a system reacts to an external force to confirm if it’s thermal <ref:2606.14542#pg3>. It suggests the response of occupation localizes and eventually saturates fundamental sensitivity bounds, which is what Aslyamov and Esposito were talking about <ref:2606.14542#pg3>.
Kai: So to summarize this first part of "Boltzmann-like occupation of nonequilibrium steady states on dense disordered networks," they’re proving that NESS on large dense networks have Boltzmann-like occupation despite the high entropy production, and the low rattling heuristic is asymptotically exact <ref:2606.14542#pg1>.
Mira: And they are tying this to equiaccessible steady states, showing that their occupation scales like e to the negative of a log of the leaving rate, which is gamma i, where gamma i is the logarithm of w i <ref:2606.14542#pg3>.
Lev: The main caveat here, and something they flag, is that they establish a nonequilibrium fluctuation-dissipation theorem that means static response isn't sufficient to confirm thermal equilibrium <ref:2606.14542#pg3>. This is important because it sets a limit on what experimental verification can prove about a steady state <ref:2606.14542#pg3>.
Kai: Moving into the conclusion of this paper, we look at the title and authors again. They’re tackling extending the Boltzmann distribution to nonequilibrium steady states, and their finding is that for large dense networks, this extension works in a specific way <ref:2606.14542#pg1>.
Mira: The implication is that we can replace more complex distributions with simpler ones like the Boltzmann one for a whole class of NESS when the network is big enough <ref:2606.14542#pg3>. It simplifies how we have to think about the steady-state occupation <ref:2606.14542#pg3>.
Lev: From an error correction view, this suggests that if you model a system as having low rattling, you’re getting a very accurate picture of the state probabilities over long times <ref:2606.14542#pg3>. But remember the paper's limitation—it relies on N being large enough for the Markov jump process to be irreducible and for pi to be unique <ref:2606.14542#pg2>.
Kai: So, what does this mean for us who only listen to the show? It means that when we look at very large, complex systems, even if they aren't perfectly thermal, the way things settle into their steady state is surprisingly close to what we expect from basic physics <ref:2606.14542#pg1>.
Mira: It gives us a simpler tool for understanding occupation in these scenarios by relating it to how slowly states are left, which is a measurable property of the state itself <ref:2606.14542#pg3>.
Lev: The future work they point toward is designing dynamics that produce these kinds of occupation patterns, showing how exit rates concentrate around their expected values and stationary probabilities become more uniform <ref:2606.14542#pg3>. That’s where the real design challenge lies for building things that exhibit these properties <ref:2606.14542#pg3>.
Kai: It’s a lot to take in, but the core idea is that simple properties of individual states can tell you a lot about the global occupation of the system when it's operating far from equilibrium <ref:2606.14542#pg3>.
Mira: Yeah, it connects microscopic state properties to macroscopic behavior in a way that extends beyond just thermal equilibrium <ref:2606.14542#pg1>.
Conclusion: Kai: So we’re wrapping up this look at "Boltzmann-like occupation of nonequilibrium steady states on dense disordered networks." Basically, they’re proving that even when things aren't in thermal equilibrium, if the network is really big and dense, the way particles settle into their steady state still looks like the standard Boltzmann distribution.
Mira: That’s actually a pretty big claim because we usually expect some kind of massive deviation from equilibrium behavior when you introduce strong driving or dissipation.
Lev: And what they’re doing is showing that even with all that entropy production happening, as the network size goes up, the difference between the actual steady state and the ideal equilibrium distribution shrinks to zero.
Kai: So it’s not just a small correction; it’s this convergence across all those metrics like relative entropy and total variation distance. That means we can replace a much more complicated model with something simpler when dealing with really large systems.
Mira: I see that as simplifying the theory, but I wonder if they fully capture the physics of *why* this happens under non-equilibrium conditions.
Lev: The scaling of the entropy production is also key here; it scales like N squared over Z, which shows that even though there’s a lot happening, it doesn't blow up in a way that completely invalidates this result for large N.
Kai: It suggests that the active matter heuristic we use—the idea of low rattling—is actually asymptotically exact for these steady states. That means the simple way we think about how often things jump around is really what’s happening deep down.
Mira: So it connects the microscopic state properties, like how slowly a state leaves, to a macroscopic pattern of occupation in a way that doesn't always work for standard thermal systems.
Lev: The real limit they set is with that nonequilibrium fluctuation-dissipation theorem, which tells us that just measuring how fast the occupation changes when you nudge the system isn't enough to tell if it’s truly thermal.
Kai: It means experimental verification of static response alone won't be enough to prove thermal equilibrium for these kinds of steady states. It points toward a deeper way we need to look at things.
Jacob Calvert, 2, Institute for Data Engineering and Science, Georgia Institute of Technology, Santa Fe Institute, Department of Mathematics, London School of Economics
Georgia Institute of Technology · Santa Fe Institute · London School of Economics
cond-mat.stat-mech, math-ph, math.MP, math.PR
Submitted: 2026-06-12
Updated: 2026-10-27
Comments: 10 pages, 3 figures, accepted for publication in Physical Review E
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 86/100
The gist: Boltzmann-like occupation of nonequilibrium steady states on dense networks proves that NESS on large dense networks have Boltzmann-like occupation despite extensive entropy production.
Key concepts
- Boltzmann-like occupation
- This means the probability of a state in a nonequilibrium steady state (NESS) on a large network approaches the form of a standard Boltzmann distribution, even though the system is not in thermal equilibrium. This happens when the number of states is sufficiently large.
- Nonequilibrium Steady States (NESS)
- These are stable states reached by systems that are not in thermal equilibrium, such as active matter. The paper models these as Markov jump processes where transitions between discrete states occur at specific rates defined by energy differences.
- Low Rattling
- This is a heuristic suggesting that for NESS on dense networks, the system's dynamics are dominated by small fluctuations or 'rattling' around an average state. The analysis proves this principle becomes asymptotically exact as the network size grows.
- Nonequilibrium Fluctuation-Dissipation Theorem (FDT)
- This theorem shows that simply measuring how a steady state responds to energy changes is not enough to tell if it is truly in thermal equilibrium. The response of occupation probabilities only matches equilibrium predictions up to a small error term ($o(1)$).
Terminology
Summary
Boltzmann-like occupation of nonequilibrium steady states on dense networks proves that NESS on large dense networks have Boltzmann-like occupation despite extensive entropy production. p. 0 Boltzmann-Like Occupation of Nonequilibrium Steady States on Dense Networks
Jacob Calvert1, 2, 3, ∗
1
Institute for Data Engineering and Science, Georgia Institute of Technology, Atlanta, GA, USA
2
Santa Fe Institute, Santa Fe, NM, USA
3Department of Mathematics, London School of Economics, London, UK
(Dated: June 15, 2026)
A central problem in statistical physics is to extend the Boltzmann distribution to nonequilibrium steady
states (NESS). We prove that NESS on large dense networks have Boltzmann-like occupation despite extensive
entropy production. We prove that NESS on large dense networks have Boltzmann-like occupation despite extensive
entropy production. We further show that the active-matter heuristic of “low rattling” is asymptotically exact.Where the paper says it
Where the paper says it
Modeling Nonequilibrium Steady States
The study models the steady state as a Markov jump process on discrete states where the transition rate from state j to state i is given by W ij = exp(E j X ij). The authors define the column sums of W = (W ij) i≠j as-W jj = w j = sum i≠j W ij. p. 0 For convenience, we set 𝑋𝑖𝑖 = 0. To make the column sums of 𝑾 = (𝑊𝑖𝑗)𝑖≠𝑗 equal to zero, we
define
−𝑊𝑗 𝑗 = 𝑤𝑗 = ∑
𝑖(≠𝑗)
𝑊𝑖𝑗. edu
When 𝑾 is irreducible, it has a unique stationary distribution
𝝅, defined by 𝑾 𝝅 = 𝟎 and 𝟏
⊤𝝅 = 1, which further satisfies
𝜋𝑖 > 0 [28]. When 𝑿 is symmetric, 𝝅 is the Boltzmann-like
distribution
𝜋
eq
𝑖 =
𝑒
−𝐸𝑖
𝑍
, (3)
where the vertex parameters play the role of dimensionless
energies in the partition function 𝑍 = ∑𝑖
𝑒
−𝐸𝑖
.Where the paper says it
Where the paper says it
Main Results on Occupation and Entropy Production
The main result shows that NESS on dense networks have Boltzmann-like occupation probabilities when the state space is large enough, specifically max i pi i / pi eq i - 1 → 0 as N → ∞. This convergence implies that the relative entropy and total variation distance between pi and pi eq tend to zero, along with all other Renyi and f-divergences. p. 1 The convergence in Eq. (5) is rather strong;
it implies that the relative entropy and total variation distance between 𝝅 and 𝝅
eq tend to zero, along with all other
Renyi and ´ 𝑓 -divergences [38]. For example, recall
that the EP associated with a distribution 𝒑 on the states is
defined as
𝜎(𝒑) = ∑
𝑖≠𝑗
𝑊𝑖𝑗𝑝𝑗
log 𝑊𝑖𝑗𝑝𝑗
𝑊𝑗𝑖𝑝𝑖
. (7)
We will show that 𝜎NESS = 𝜎(𝝅) scales like 𝑁
2/𝑍 [39].Where the paper says it
Where the paper says it
Equiaccessibility and Low Rattling
The analysis introduces the concept of equiaccessible steady states, which are those with uniform pi hat, whose occupation is determined by rattling i. The principle of low rattling is shown to be asymptotically exact for NESS on dense networks, as the correlation rho ≈ 1 with respect to the uniformly random choice of state. p. 2 Equation (16) implies that
it satisfies
𝜌 = Corr(𝒖 + 𝑜(1), 𝒖) → 1, (18)
for any vertex parameters satisfying Var(𝐸𝒖) ̸→ 0. In particular, the principle of low rattling as formulated in [19] is
asymptotically exact for NESS on dense networks (Fig.Where the paper says it
Nonequilibrium Fluctuation-Dissipation Theorem
The study derives a nonequilibrium FDT which implies that static response to energy perturbations is insufficient to determine whether a steady state is in thermal equilibrium. Specifically, the response of occupation to perturbations of vertex parameters satisfies dE i pi = dE i pi eq(1 + o(1)). p. 3 Fluctuation–dissipation theorem—As a further application
of Eq. (5), consider the response of NESS occupation to perturbations of the vertex parameters.Where the paper says it
Implications and Applications
The results justify replacing pi by a simpler distribution for a class of NESS beyond the near-equilibrium regime. The authors also establish that such NESS cannot satisfy a principle of near-minimal EP, as was recently conjectured. p. 3 First, we established
that the active-matter heuristic of low rattling is asymptotically exact for NESS on dense networks [18, 19] (Fig. Second, we argued that such NESS cannot satisfy a principle of
near-minimal EP, as was recently conjectured [20] (Fig.Where the paper says it
Entropy Production Scaling
The entropy production scales as sigma NESS = sigma(pi eq)(1 + o(1)). This scaling is derived by analyzing the difference between the EP of NESS and equilibrium states. p. 4 5
Third, we derived a nonequilibrium FDT, which implies that
static response to energy perturbations is insufficient to determine whether a steady state is in thermal equilibrium. Although we have emphasized what our results mean for
steady states in general, one can use the ideas of this Letter to establish many other interesting properties of NESS on
dense networks. (23)
implies that the expected value of 𝜎NESS is greater than that
of 𝜎min by a factor of order 𝑁/ log𝑁 (Fig. This example demonstrates why a principle of near-minimal EP cannot
generally hold for NESS on dense networks: The nearness of
𝜎NESS and 𝜎min is determined by the way the vertex parameters vary.Where the paper says it
Where the paper says it
Experimental Verification Limits
The derivation of the nonequilibrium FDT suggests that experimental verification of the static FDT does not suffice to conclude that a steady state is in thermal equilibrium. This strengthens a result of [25]. p. 1 , with a probability of 1 over
the randomness of the edge weights 𝑿. Main result—We find that NESS on dense networks have
Boltzmann-like occupation probabilities when the state space
is large enough. (5) to show that the response of their
occupation to perturbations of transition rates localizes. As
a consequence, they asymptotically saturate the fundamental sensitivity bounds of Aslyamov and Esposito [34, Eq.Where the paper says it
Where the paper says it
Future Directions
The work focuses on explaining occupation by showing how exit rates concentrate around their expected values and stationary probabilities become increasingly uniform. Future studies could explore designing dynamics to produce desirable occupation. p. 4 We focused on the “forward” direction of this relationship,
which entails explaining occupation, as we did for the principle of low rattling. Future studies could explore the reverse
direction, which entails designing or modifying dynamics to
produce desirable occupation. raised the possibility of a nonequilibrium analogue of the
Boltzmann distribution for precisely this reason [2]. The virtue of our results about NESS on large dense networks, and equiaccessible steady states more broadly, is that
they relate simple properties of states to global occupation.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Advanced prediction of steady-state behavior in large dense networks: The system can accurately infer steady-state distributions from local measurements made away from stationarity because
the steady-state distributions of such NESS can be inferred through local measurements made away from stationarity.
-
Robust analysis of active matter dynamics: The AI can apply the active matter heuristic of
low rattling
to predict occupation patterns, as this heuristic is shown to beasymptotically exact for such NESS.
-
Characterization of entropy production in complex systems: The system can estimate the entropy production (EP) for nonequilibrium steady states by showing that
sigma NESS scales like N 2/Z,
which allows it to quantify how extensive EP coexists with Boltzmann-like occupation. -
Understanding response to external perturbations: The AI can predict the response of state occupation to changes in vertex parameters, as the results show that
NESS on dense networks in particular satisfy dEpi = dEpi eq(1 + o(1)).
-
Verification of thermal equilibrium conditions: The system can determine if a steady state is truly thermal by using a nonequilibrium fluctuation-dissipation theorem, predicting that
experimental verification of the value of the static FDT does not suffice to conclude that a steady state is in thermal equilibrium.
Abstract
A central problem in statistical physics is to extend the Boltzmann distribution to nonequilibrium steady states (NESS). We prove that NESS on large dense disordered networks have Boltzmann-like occupation despite extensive entropy production. One consequence is that such NESS can conceal their nonequilibrium nature from measurements of static response to energy perturbations. We further prove that the active-matter heuristic of "low rattling" is asymptotically exact for these NESS. Intuitively, they spend a greater fraction of their time in states they leave more slowly. We show that rattling effectively extends the Boltzmann distribution to such NESS, as well as the broader class of "equiaccessible" steady states, which play a role in our analysis akin to that of equilibrium in linear response.
Sources
- Large Interconnected Thermodynamic Systems Nearly Minimize Entropy Production
- Fast and Slow Mixing of the Kawasaki Dynamics on Bounded-Degree Graphs
Related papers
- Emergent long-tail dynamics in driven magnets with dynamical frustration
- Statistical Mechanics of Dynamical System Identification
- Second law of thermodynamics in closed quantum many-body systems
- Chemical potentials from structure factors: II. Charged multi-component mixtures
- Genuine quantum scars in Floquet chaotic many-body systems
- Stochastic Thermodynamics for Autoregressive Generative Models: A Non-Markovian Perspective