🔢 mathematics

Parity Selection Rule for Information and Dissipation in Driven Steady States

This paper establishes a parity selection rule in rotation-driven linear nonequilibrium steady states that forbids tight equalities between symmetric information and entropy production by demonstrating that mutual information is strictly even under drive reversal while entropy production is quadratic, a phenomenon that persists even under heavy-tailed stable noise.

Original authors: Mengqi Li, Lixin Li, Wensheng Lin, Zhu Han

Published 2026-06-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mengqi Li, Lixin Li, Wensheng Lin, Zhu Han

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 watching a tiny, invisible particle dancing in a fluid. Sometimes, you push it with a gentle, spinning wind (a "drive"). Sometimes, the fluid is calm. Scientists have long wondered if there is a perfect, one-to-one rule connecting how much the particle moves (information) with how much energy is wasted as heat (dissipation). They hoped to find a simple equation where "Information = Heat."

This paper says: No, that perfect equation doesn't exist. In fact, the universe has a hidden "traffic rule" that makes it impossible for these two things to ever be exactly equal in a spinning system.

Here is the breakdown using simple analogies:

1. The Two Types of "Dance Moves" (Even vs. Odd)

The authors discovered that in a spinning system, everything falls into one of two categories based on how it reacts if you reverse the direction of the spin (like playing a video backward):

The "Even" Dancers (Information): Imagine a photo of the particle's position. If you spin the wind clockwise or counter-clockwise, the relationship between where the particle was a moment ago and where it is now looks exactly the same. It doesn't care which way you spin. This is "Even."
- The Analogy: Think of a perfect circle. If you rotate it left or right, it still looks like a circle. The "snapshot mutual information" (how much you know about the past based on the present) is like this circle. It stays constant regardless of the spin speed.
The "Odd" Dancers (Heat/Dissipation): Now imagine the heat being generated. If you spin clockwise, you generate heat. If you spin counter-clockwise, you also generate heat. But mathematically, the "direction" of the heat flow flips. The amount of heat grows with the square of the spin speed.
- The Analogy: Think of a car engine. Whether you drive forward or backward, the engine gets hot. But the "heat" is tied to the speed squared. It's an "Odd" function because the underlying physics flips sign when you reverse the drive, even though the heat amount stays positive.

2. The "Parity Selection Rule" (The Traffic Light)

The paper calls this the Parity Selection Rule. It's like a strict traffic light that says: "You cannot mix these two types of dancers."

Because "Information" is an Even dancer (it stays the same when you reverse the spin) and "Heat" is an Odd dancer (it flips its mathematical sign when you reverse the spin), they can never be equal to each other.

If you try to write an equation saying Information = Heat, the math breaks. It's like trying to say Left Hand = Right Hand when one is a mirror image of the other.
The only thing you can say is a "one-sided" rule (an inequality): "Heat must be at least this much," but you can never say "Heat is exactly this much information."

3. The "Blind Spot" (Isotropy)

The paper finds a special case where the system is perfectly symmetrical (like a perfect sphere). In this case:

The Information becomes completely "blind" to the spin. No matter how fast you spin the wind, the information stays at a fixed, constant number (about 0.145 "nats," a unit of information).
The Heat, however, keeps growing as you spin faster.
The Result: As you spin slower and slower, the heat drops to zero, but the information stays at that constant number. They drift infinitely far apart. This proves they can never be the same thing.

4. What if the Noise is "Weird"? (Heavy-Tailed Noise)

Usually, scientists assume the fluid particles bump into each other in a standard, predictable way (Gaussian noise). But what if the bumps are wild and unpredictable, with occasional massive jumps (like a storm)?

Standard rules for heat and information break down here because the math involves "infinite variance" (the jumps are too wild to measure with a ruler).
The Paper's Big Surprise: The Parity Selection Rule still works! Even in this chaotic, wild environment, the "Even" nature of the information and the "Odd" nature of the heat remain true. The rule survives because it depends on the symmetry of the spin, not on the specific size of the bumps.

5. The Proposed Experiment

The authors suggest testing this with a real electronic circuit called a Brownian Gyrator.

Imagine two capacitors (electrical buckets) connected by resistors.
By heating one side more than the other, you create a "wind" that makes the electrical signals spin in a circle.
They propose adding a special noise generator to create those "wild jumps" (heavy-tailed noise) to see if the rule holds up in the real world.

Summary

The paper argues that nature has a fundamental symmetry that prevents us from finding a perfect, tight equation linking information and heat in spinning systems.

Information is like a static photo; it doesn't care about the direction of the spin.
Heat is like a spinning top; it depends heavily on the speed and direction.
Because they behave so differently, you can never say "Information = Heat." You can only say "Heat is at least this much."

This isn't just a math trick; it's a fundamental law of how driven systems work, surviving even in the most chaotic, unpredictable environments.

Technical Summary: Parity Selection Rule for Information and Dissipation in Driven Steady States

Problem Statement
The paper addresses a fundamental gap in stochastic thermodynamics: the inability to establish tight equalities between symmetric information measures and entropy production in driven nonequilibrium steady states (NESS). While fluctuation theorems and entropy-production identities provide a robust framework, and recent work has established tight equalities for time-antisymmetric information measures, no such equality exists for the natural symmetric information content of a steady state. Previous interpretations suggested this was a missing construction; this work argues instead that the obstruction is a fundamental symmetry constraint.

Methodology
The authors analyze a rotation-driven linear nonequilibrium steady state modeled by an $n$ -dimensional Ornstein-Uhlenbeck (OU) process:
$d\mathbf{x}_t = -(A - \Omega)\mathbf{x}_t dt + d\mathbf{N}_t$
where $A$ is a symmetric positive-definite relaxation matrix, $\Omega$ is an antisymmetric rotational drive matrix, and $\mathbf{N}_t$ represents noise. The study focuses on the planar case ( $n=2$ ) and investigates the behavior of steady-state functionals under drive reversal ( $\omega \to -\omega$ ).

The analysis considers two noise regimes:

Gaussian: Standard Brownian motion with finite variance.
Heavy-tailed: Isotropic $\alpha$ -stable processes ( $\gamma \in (0, 2)$ ) where variance is infinite, rendering standard variance-based thermodynamic uncertainty relations (TURs) vacuous.

The core methodological tool is the parity selection rule, which categorizes steady-state functionals into "even" and "odd" sectors based on their transformation properties under drive reversal. The authors specifically examine the structural alignment condition $[A, D] = 0$ (where $D$ is the diffusion matrix) and the stronger condition of full isotropy ( $A = aI, D = D_0I$ ).

Key Contributions and Results

The Parity Selection Rule:
The paper establishes that for rotation-driven linear NESS with aligned relaxation and diffusion matrices ( $[A, D] = 0$ ):
- Even Branch (Symmetric Information): The snapshot mutual information $I(\mathbf{x}_0; \mathbf{x}_t)$ between two time slices is exactly even under drive reversal ( $\omega \to -\omega$ ). Under full isotropy, this mutual information becomes completely independent of the drive frequency $\omega$ , taking a closed-form value of approximately $0.145$ nats for the planar case ( $n=2$ ).
- Odd Branch (Entropy Production): The entropy production rate $\Sigma$ is odd-rooted. Its directional content is time-antisymmetric, and under drive reversal, it flips sign. Under isotropy, $\Sigma$ is exactly quadratic in the drive ( $\Sigma \propto \omega^2$ ).
The "No-Go" for Tight Equalities:
Because the symmetric information functional resides in the even sector and the entropy production resides in the odd sector (specifically scaling as $\omega^2$ vs. $\omega^0$ ), the authors prove that no non-trivial tight equality of the form $I(\omega) = \Sigma(\omega)$ can hold for all $\omega$ . The orthogonality of even and odd sectors forces any such identity to imply $I \equiv 0$ , which contradicts the non-triviality of the information content. This explains why previous attempts to find such equalities failed; the obstruction is symmetry-based, not a lack of construction.
Robustness in Heavy-Tailed Regimes:
A significant contribution is the extension of this rule to the heavy-tailed regime ( $\gamma < 2$ ). While variance-based TURs break down here (due to infinite variance of circulation observables), the parity selection rule remains exact. This is because the rule relies on the rotational symmetry of the drift term rather than the moments of the noise. The authors demonstrate that the mutual information remains finite and $\omega$ -independent even when the stationary covariance is undefined, provided the noise is isotropic $\alpha$ -stable.
One-Sided Bounds:
The rule does not forbid one-sided inequalities (such as those found in thermodynamic uncertainty relations) but strictly forbids the "tight merger" of information and dissipation. The ratio $I/\Sigma$ diverges as $\omega \to 0$ , confirming that dissipation (scaling as $\omega^2$ ) cannot bound information (scaling as $\omega^0$ ) with a finite constant in the low-drive limit.

Experimental Proposal
The paper proposes a falsifiable test using an electrical autonomous Brownian gyrator (two capacitively coupled RC branches with a temperature gradient).

Gaussian Test: Verify the $\omega$ -independence of mutual information at the symmetric point ( $R_1 \approx R_2$ ) and the quadratic scaling of entropy production.
Heavy-Tailed Test: Augment the circuit with a programmable current source to inject isotropic $\alpha$ -stable noise. The proposal requires an independent control mechanism for the antisymmetric drive $\Omega$ (e.g., via non-reciprocal coupling or active feedback) to decouple the drive from the diffusion matrix $D$ . This setup allows testing the parity rule in the regime where variance-based bounds are vacuous.

Significance
The paper claims to resolve the "missing construction" problem by identifying a fundamental symmetry obstruction. It reframes the relationship between information and dissipation:

It clarifies that the symmetric information content is "blind" to the drive direction under specific symmetries, while dissipation is "sensitive" to it.
It establishes that the failure to equate information and entropy production is not a technical limitation but a consequence of the parity selection rule.
It provides a symmetry-based framework that survives beyond the Gaussian regime, offering a more robust characterization of nonequilibrium steady states than variance-based methods.

The work distinguishes itself from concurrent research by focusing on the "even" (symmetric) information half, whereas other recent works have focused on the "odd" (directed) information half or different nonequilibrium objects. The authors position their result as a companion to recent findings on dissipation constraints, together forming a complete symmetry-based picture of NESS functionals.