{
 "artifacts": null,
 "category": "dynamics",
 "description": "Train two parameter replicas with common low-rank stochastic forcing and an adaptive finite-dimensional Cameron–Martin correction that contracts their discrepancy in a weak parameter metric. Transporting the forcing directions through the loss Hessian is intended to make a rank-k perturbation influence more than k raw parameter directions, while damped momentum suppresses high-energy divergence.",
 "formulas_latex": [
  "$$\\mathrm{d}u=v\\,\\mathrm{d}t,\\qquad\\mathrm{d}v=(\\Delta u-a(x)v-u^{3})\\,\\mathrm{d}t+B\\,\\mathrm{d}W_t.$$",
  "$$\\mathcal{E}_{s}=H^{-s}(\\mathbb{T}^{3})\\times H^{-1-s}(\\mathbb{T}^{3}),\\qquad 0\u003cs\u003c\\tfrac12.$$",
  "$$Q_r=[B\\; HB\\; H^2B\\;\\cdots\\;H^{r-1}B],\\qquad \\rho_r=\\frac{\\operatorname{rank}_\\varepsilon(Q_r)}{d}.$$",
  "$$\\mathrm{d}\\theta=p\\,\\mathrm{d}t,\\qquad \\mathrm{d}p=\\bigl(-\\nabla L(\\theta)-\\gamma p\\bigr)\\,\\mathrm{d}t+B\\,\\mathrm{d}W_t,\\qquad \\gamma\u003e0.$$"
 ],
 "id": 2942,
 "implementation": "(1) Integration point: replace the momentum update in SGD with a two-replica optimizer. Maintain $(\\theta^A,p^A)$ and $(\\theta^B,p^B)$, use the same minibatches and Gaussian vector $\\xi_t\\in\\mathbb{R}^k$ for both replicas, and inject $B\\xi_t$ only into momentum. Choose $B$ as k normalized random parameter directions or low-rank directions from recent gradient covariance. Define a weak discrepancy $D_w=\\|P_m(\\theta^A-\\theta^B)\\|_2+\\alpha\\|P_m(p^A-p^B)\\|_2+\\beta\\|(I-P_m)(\\theta^A-\\theta^B)\\|_2/\\sqrt{m}$, where $P_m$ retains low-frequency or randomly selected parameter blocks and therefore downweights fine-scale discrepancies. (2) Pseudocode: compute paired gradients $g_A,g_B$; update $p^i\\leftarrow(1-\\gamma\\eta)p^i-\\eta g_i+\\sigma\\sqrt{\\eta}B\\xi$ and $\\theta^i\\leftarrow\\theta^i+\\eta p^i$. Every T steps, compute Hessian-vector products at $\\bar\\theta=(\\theta^A+\\theta^B)/2$, append $B,HB,\\ldots,H^{r-1}B$ to $Q_r$, and compute a thin SVD. If $D_w$ is increasing, apply $c=Q_r(Q_r^TQ_r+\\lambda I)^{-1}(\\theta^B-\\theta^A)$ and set $\\theta^B\\leftarrow\\theta^B+\\kappa c$, with $\\kappa$ clipped to $[-\\kappa_{max},\\kappa_{max}]$. (3) Compute from the paper's mechanism: damped second-order dynamics, finite-rank forcing, weak metric, and finite-dimensional correction. Estimate empirically: Hessian-vector products, $\\rho_r$, correction coefficients, and contraction rates; never form H explicitly. (4) First experiment: train a two-layer width-512 MLP on MNIST or CIFAR-10 against SGD with momentum, AdamW, and isotropic underdamped Langevin noise. Match parameter count, optimizer steps, batch size, and total HVP FLOPs. Sweep $k\\in\\{1,2,4,8,16\\}$ and $r\\in\\{1,2,4,8\\}$. (5) Pre-register predictions: improvement should appear when $\\rho_r$ exceeds 0.5, and the slope of $\\log D_w$ over 200-step windows should correlate negatively with $\\rho_r$ (Spearman correlation below -0.5). Correction should help mostly below the 90th percentile of $\\|p\\|^2+\\|g\\|^2$, while damping should dominate above it. Require 20% faster loss decrease at equal FLOPs or 10% lower seed variance. Ablate only transport by replacing $HB,H^2B$ with fresh random directions while preserving rank, noise variance, damping, and correction frequency. Falsify the transfer if transported directions do not outperform fresh random directions, if $\\rho_r$ fails to predict contraction, or if gains vanish after HVP costs are equalized.",
 "math_summary": "The source equation is the damped second-order stochastic system $\\mathrm{d}u=v\\,\\mathrm{d}t$, $\\mathrm{d}v=(\\Delta u-a(x)v-u^{3})\\,\\mathrm{d}t+B\\,\\mathrm{d}W_t$, where $u$ is a field, $v$ is its velocity, $a(x)\u003e0$ is damping, $B$ maps finite-dimensional Brownian motion into the state space, and $W_t$ is Brownian motion. The paper measures phase discrepancies in the weak negative topology $\\mathcal{E}_s=H^{-s}(\\mathbb{T}^3)\\times H^{-1-s}(\\mathbb{T}^3)$ for $0\u003cs\u003c1/2$, rather than the strong energy topology. Its stable–compact asymptotic coupling principle separates contraction into a compact defect in the weak topology plus a finite-dimensional Cameron–Martin shift, with dissipation handling high-energy states. The ML adaptation uses parameters and momentum $(\\theta,p)$, injects noise through $B\\in\\mathbb{R}^{d\\times k}$ with $k\\ll d$, and estimates nonlinear saturation using the empirical span of $B,HB,H^2B,\\ldots$, where $H$ is a Hessian-vector-product approximation. Define $Q_r=[B,HB,\\ldots,H^{r-1}B]$ and $\\rho_r=\\operatorname{rank}_\\varepsilon(Q_r)/d$, where $\\operatorname{rank}_\\varepsilon$ counts singular values above tolerance $\\varepsilon$.",
 "math_tags": [
  "dynamical-systems",
  "stochastic-processes",
  "control-theory",
  "geometry"
 ],
 "ml_areas": [
  "optimizer",
  "training-dynamics",
  "regularization"
 ],
 "paper": {
  "arxiv_id": "2608.28459",
  "arxiv_url": "https://arxiv.org/abs/2608.28459",
  "summary_what_math_gives_to_ml": "The paper develops a non-parabolic mixing mechanism in which finite-rank noise becomes effective through nonlinear saturation: a finite-dimensional Cameron–Martin correction compensates a compact mismatch in a deliberately weaker topology, while dissipation controls large-energy states. The transferable asset is not the wave equation itself, but the separation of contraction into a weak-topology compact-defect step and a high-energy dissipative step. A practical experiment can test whether low-rank stochastic control directions become broadly effective after transport through network curvature, improving stability or escape from bad basins at fixed compute.",
  "title": "Spectral gap for the three-dimensional damped cubic wave equation with degenerate noise",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 7,
  "novelty": 7,
  "usefulness": 5
 },
 "solves": [
  "stability",
  "generalization",
  "sample-efficiency"
 ],
 "title": "Saturating low-rank coupled optimizer",
 "url": "https://synthcore.org/idea/2942/saturating-low-rank-coupled-optimizer",
 "verification": {
  "peer_reviewed": false,
  "status": "unverified",
  "status_label": "Unverified",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)",
  "verification_axes": {
   "benchmark_mechanism": {
    "confirmed": null,
    "tested": false
   },
   "practical_benchmark": {
    "beats_baseline": null,
    "tested": false
   },
   "toy_mechanism_gate": {
    "confirmed": null,
    "tested": false
   }
  }
 }
}