{
 "artifacts": null,
 "category": "dynamics",
 "description": "Treat a momentum optimizer or recurrent state update as a damped oscillatory feedback system whose local closed-loop dynamics have a cubic characteristic polynomial. Estimate local damping, oscillation frequency, and feedback gain, then cap the learning-rate or momentum gain using the cubic Routh-Hurwitz inequality so that oscillatory divergence is prevented before it appears in the loss.",
 "download_zip": "https://synthcore.org/download/1165",
 "formulas_latex": [
  "p(s) = |sI − A| = s³ + (2r/L)s² + (r²/L² + ω₀²)s + (3/2)(ω₀/L)(mₚ₁V₁₀² + mₚ₂V₂₀²)",
  "a₁ \u003e 0, a₂ \u003e 0, a₃ \u003e 0, and a₁a₂ \u003e a₃",
  "g_max = ρ â₁â₂/κ, where â₁ = 2r̂/L̂, â₂ = r̂²/L̂² + ω̂₀², and κ = (3/2)(ω̂₀/L̂)",
  "χ = a₃/(a₁a₂); χ \u003c 1 predicts continuous-time stability, while χ approaching 1 from below predicts a pole-crossing boundary"
 ],
 "id": 3012,
 "implementation": "1. Integration point: implement this as a wrapper around SGD with momentum, Adam without its elementwise preconditioner, or a recurrent state-space update. Maintain a scalar effective feedback gain g representing learning rate multiplied by local curvature or feedback amplification; for a layerwise version, compute one cap per parameter block. Use the previous two or three parameter and gradient states to estimate local oscillatory dynamics. 2. Pseudocode: every K optimization steps, fit a least-squares local model z(t+1) = B₁z(t) + B₂z(t−1) + B₃z(t−2), or estimate the dominant complex eigenpair using Jacobian-vector-product probes. Convert the dominant pair into r̂, L̂, and ω̂₀; alternatively fit cubic coefficients directly and obtain â₁ and â₂ from the fitted polynomial. Compute g_max = ρâ₁â₂/κ with ρ = 0.8, then set g to the minimum of its proposed value and g_max before applying the optimizer update. Log χ = â₃/(â₁â₂). 3. Computed versus estimated: the cubic Routh-Hurwitz inequality is the paper-derived condition; all neural quantities are empirical local estimates and should be smoothed with an exponential moving average. Reject estimates with negative damping or poor fit residuals and fall back to a conservative fixed learning rate. 4. First cheap experiment: train a two-layer MLP on MNIST with SGD momentum while sweeping learning rate and momentum. Compare vanilla SGD, gradient clipping, and this cap. For each run measure loss variance, oscillation amplitude, fitted cubic roots, and χ. The falsifiable prediction is a sharp rise in oscillation amplitude and loss variance as χ approaches one, with the divergence boundary predicted by χ = 1 within approximately 20 percent. The capped method should keep χ below 0.8 and eliminate runs that diverge at the same nominal learning rate.",
 "math_summary": "The paper obtains the characteristic polynomial p(s) = s³ + a₁s² + a₂s + a₃, where a₁ = 2r/L, a₂ = r²/L² + ω₀², and a₃ = (3/2)(ω₀/L)(mₚ₁V₁₀² + mₚ₂V₂₀²). For a real cubic, asymptotic stability requires a₁ \u003e 0, a₂ \u003e 0, a₃ \u003e 0, and a₁a₂ \u003e a₃. Here r \u003e 0 is damping, L \u003e 0 is an inductive or time-scale parameter, ω₀ \u003e 0 is the nominal oscillation frequency, mₚ₁ and mₚ₂ are feedback gains, and V₁₀ and V₂₀ are operating-point amplitudes. In a neural optimizer, define an estimated effective feedback gain g such that a₃ = κg, with κ = (3/2)(ω̂₀/L̂). Enforce g \u003c ρâ₁â₂/κ with safety factor ρ \u003c 1. The mechanism predicts instability when χ = a₃/(a₁a₂) reaches one.",
 "math_tags": [
  "control-theory",
  "dynamical-systems",
  "spectral-theory",
  "linear-algebra",
  "optimization"
 ],
 "ml_areas": [
  "optimizer",
  "training-dynamics",
  "scheduler",
  "rnn",
  "ssm"
 ],
 "paper": {
  "arxiv_id": "2609.01571",
  "arxiv_url": "https://arxiv.org/abs/2609.01571",
  "summary_what_math_gives_to_ml": "The paper derives a cubic characteristic polynomial for a two-converter low-frequency transmission-line system with power-frequency droop gains and uses pole migration to determine stability. Its transferable mechanism is a quantitative gain ceiling: for a cubic polynomial, the Routh-Hurwitz condition imposes an upper bound on effective feedback gain, rather than merely requiring positive damping. This can be transferred to neural-network optimizers or recurrent state updates by estimating the local damped-oscillatory Jacobian and clipping the learning or feedback gain before the discrete dynamics cross their stability boundary.",
  "title": "Stability of Droop-Controlled Low-Frequency Transmission Lines",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 6,
  "novelty": 7,
  "usefulness": 7
 },
 "solves": [
  "stability",
  "speed"
 ],
 "title": "Routh-Hurwitz Gain-Capped Optimizer",
 "url": "https://synthcore.org/idea/3012/routh-hurwitz-gain-capped-optimizer",
 "verification": {
  "peer_reviewed": false,
  "status": "queued_mechanism",
  "status_label": "Queued — mechanism check",
  "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
   }
  }
 }
}