# Conservative Chapman–Enskog Neural Layer

- ID: 1067
- Canonical URL: https://synthcore.org/idea/1067/conservative-chapman-enskog-neural-layer
- API JSON: https://synthcore.org/api/idea/1067.json
- API Markdown: https://synthcore.org/api/idea/1067.md
- Verification status: beats_baseline
- Source: [arXiv:2607.17358](https://arxiv.org/abs/2607.17358)
- Category: architecture
- Solves: stability, accuracy, memory
- ML areas: rnn, ssm, training-dynamics
- Math tags: dynamical-systems, control-theory, spectral-theory, linear-algebra
- Ratings: usefulness 8/10; difficulty 7/10; novelty 7/10

## Idea description

Replace an unconstrained recurrent hidden-state update by a fast redistribution state with a dissipative Jacobian and a slow conserved state. The network computes an equilibrium state and a first-order pseudoinverse response correction, transferring the paper’s separation between local relaxation and macroscopic transport into a stable recurrent or state-space layer.

## Mathematical statement

Let $g\in\mathbb{R}^n$ be a fast hidden occupancy vector, $m=w^\top g$ the conserved scalar state, and $C(g)$ a redistribution vector field satisfying $w^\top C(g)=0$. At an equilibrium $g_{\rm eq}(m)$, define $J=\partial C/\partial g|_{g_{\rm eq}}$. The mechanism requires $J$ to be self-adjoint in a positive mass-weighted inner product and negative semidefinite, with $\ker J=\mathrm{span}\{w\}$, so only the component orthogonal to the conserved mode relaxes. For a slowly varying forcing $S$, the first Chapman–Enskog correction is $\chi=-J^\dagger S$, where $J^\dagger$ is the pseudoinverse on the decaying subspace. The spectral gap $\gamma=-\max\{\lambda<0:\lambda\in\mathrm{spec}(J)\}$ controls the fast decay and the validity of the reduction.

## Key formulas

- $$\frac{\partial g}{\partial t}=\frac{1}{\mathrm{Da}}C_\theta(g)+S_\theta(x),\qquad m=w^\top g,\qquad w^\top C_\theta(g)=0.$$
- $$J_\theta=\left.\frac{\partial C_\theta}{\partial g}\right|_{g_{\rm eq}},\qquad J_\theta=J_\theta^\top,\qquad u^\top J_\theta u\le 0,\qquad \ker J_\theta=\mathrm{span}\{w\}.$$
- $$g(t)=g_{\rm eq}(m(t))+\mathrm{Da}\,\chi(t)+O(\mathrm{Da}^2),\qquad \chi=-J_\theta^\dagger S_\theta.$$
- $$\dot V=\frac{1}{\mathrm{Da}}(g-g_{\rm eq})^\top J_\theta(g-g_{\rm eq})\le-\frac{\gamma_\theta}{\mathrm{Da}}\|g-g_{\rm eq}\|^2.$$

## Implementation notes

Use this as a recurrent cell or continuous-time state-space layer. Let an input projection produce a scalar or low-dimensional conserved state $m_t$ and parameters $\theta_t$. Define a redistribution field $C_\theta(g)$ with exact conservation, for example by using a graph-Laplacian parameterization whose column sums vanish. Compute $g_{\rm eq}(m_t,\theta_t)$ by several fixed-point iterations while enforcing $w^\top g_{\rm eq}=m_t$. Use automatic differentiation to estimate the Jacobian $J$ at equilibrium, project out the conserved direction with $P=I-w(w^\top w)^{-1}w^\top$, and solve $PJP\chi=-PS$ by conjugate gradients instead of explicitly forming an inverse. Update the fast state using $g_{t+1}=g_{\rm eq}+\mathrm{Da}\chi$ or initialize an implicit Euler step with this approximation. Update the slow state as $m_{t+1}=m_t+\Delta t\,w^\top S$. The paper-derived ingredients are the conserved nullspace, dissipative Jacobian, and pseudoinverse response; the network estimates $J$, equilibrium, and forcing from data. The first cheap experiment is a GRU replacement on sequential MNIST and Copy Memory, compared with a GRU and implicit RNN at matched parameter count. Sweep $\mathrm{Da}$ and timestep $\Delta t$. Prediction: perturbations orthogonal to $w$ decay as approximately $\exp(-\gamma t/\mathrm{Da})$, while explicit integration becomes unstable near $\Delta t|\lambda_{\min}(J)|/\mathrm{Da}=2$; the measured boundary should agree within 20%.

## Verification

- Status: beats_baseline
- Mechanism evidence: yes
- Mechanism confirmed: yes
- Practical verdict: helps
- Verdict: Built a conservative recurrent layer with an exactly conserved mode and symmetric negative-semidefinite redistribution Jacobian, plus numerical verification and a sequence-sum mini-experiment. The mechanism manifested: decay-rate error was 0.50%, the explicit-Euler stability boundary matched the predicted boundary within 0.02%, and the Chapman–Enskog response scaled linearly with Da with negligible error. On the learning task, the conservative layer had essentially identical test MSE to the tanh RNN, so no meaningful task-level accuracy win was observed.

### Mechanism check

- Verdict: Built a conservative recurrent layer with an exactly conserved mode and symmetric negative-semidefinite redistribution Jacobian, plus numerical verification and a sequence-sum mini-experiment. The mechanism manifested: decay-rate error was 0.50%, the explicit-Euler stability boundary matched the predicted boundary within 0.02%, and the Chapman–Enskog response scaled linearly with Da with negligible error. On the learning task, the conservative layer had essentially identical test MSE to the tanh RNN, so no meaningful task-level accuracy win was observed.
- Confidence: 9/10
- Limitations: Only a tiny synthetic length-30 sequence-sum task was tested, not sequential MNIST, Copy Memory, GRU, or implicit Euler. The learning comparison used equal hidden width but not equal parameter count, and only one fixed seed/training configuration; the stability test used a linear fixed operator rather than a learned nonlinear equilibrium solver.

### Practical benchmark

- Paired seeds: 8
- Baseline mean: 0
- Idea mean: 0
- p-value: 0.0252
- Paired wins: 7/8
- Benchmark verdict: idea better (significant)

## Artifacts

- [bench_report.json](https://synthcore.org/code/189/bench_report.json)
- [bench_runner.py](https://synthcore.org/code/189/bench_runner.py)
- [mini_experiment.py](https://synthcore.org/code/189/mini_experiment.py)
- [mini_results.json](https://synthcore.org/code/189/mini_results.json)
- [report.md](https://synthcore.org/code/189/report.md)
- [report_bench_2026-08-30T194456.md](https://synthcore.org/code/189/report_bench_2026-08-30T194456.md)
- [results.json](https://synthcore.org/code/189/results.json)
- [verify_conservative_layer.py](https://synthcore.org/code/189/verify_conservative_layer.py)
- [Download all files as ZIP](https://synthcore.org/download/189)

## Disclaimer

AI-generated research hypothesis, automatically tested. Not peer-reviewed.