{
 "artifacts": null,
 "category": "dynamics",
 "description": "Add a bounded colored exploration force to an optimizer by filtering a sum of independent two-state telegraph signals through a stable linear relaxation equation. Unlike Gaussian momentum noise, the perturbation has a strict amplitude bound and a tunable finite correlation time, reducing rare destructive parameter excursions while retaining structured exploration.",
 "formulas_latex": [
  "du/dt = -gamma u(t) + sum_k v_k s_k(t), with s_k(t) in {-1,+1} and flip probability lambda_k dt over a small interval dt.",
  "|u(t)| \u003c= (sum_k v_k)/gamma = u_max.",
  "Var(u) = sum_k v_k^2/[gamma(gamma + 2 lambda_k)].",
  "C_u(tau) = sum_k v_k^2 [exp(-2 lambda_k |tau|) - exp(-gamma |tau|)]/[gamma^2 - 4 lambda_k^2], for gamma != 2 lambda_k."
 ],
 "id": 2817,
 "implementation": "Integrate the mechanism into a momentum-like optimizer state rather than directly perturbing weights. For parameter vector theta, maintain telegraph force u and update theta \u003c- theta - eta [g(theta) + alpha u], where g is the minibatch gradient, eta is the learning rate, and alpha scales exploration. Use one scalar telegraph process per parameter block or layer to avoid storing a sign for every parameter. With step size dt, flip each sign s_k independently with probability 1 - exp(-lambda_k dt), then apply the exact linear relaxation update u \u003c- exp(-gamma dt) u + [1 - exp(-gamma dt)] sum_k v_k s_k/gamma. Pseudocode: initialize u = 0 and random signs; for each minibatch compute g; flip signs; update u; optionally clamp only roundoff excursions to [-u_max,u_max]; apply theta \u003c- theta - eta(g + alpha u). The paper supplies the bounded-support and stationary-moment mechanism; estimate empirical variance and autocorrelation from the run. First test a small MLP and ResNet-18 on CIFAR-10 against SGD and SGD with Gaussian OU noise at matched variance. Sweep gamma, K, and lambda_k. The quantitative prediction is |u| \u003c= u_max at every step, with violations below numerical precision. Measured stationary variance should match the closed form within 10%, and the autocorrelation should fit the predicted sum of exponentials. As K increases with fixed total variance, standardized excess kurtosis should decrease toward zero.",
 "math_summary": "For each parameter coordinate or parameter block, define u(t) by du/dt = -gamma u(t) + sum_k v_k s_k(t), where gamma \u003e 0 is the relaxation rate, v_k \u003e 0 are amplitudes, and s_k(t) in {-1,+1} are independent symmetric telegraph processes switching at rate lambda_k. The stationary process obeys |u(t)| \u003c= V/gamma, where V = sum_k v_k. For symmetric switching, E[u] = 0, Var(u) = sum_k v_k^2/[gamma(gamma+2 lambda_k)], and the autocovariance is C_u(tau) = sum_k v_k^2[exp(-2 lambda_k |tau|) - exp(-gamma |tau|)]/[gamma^2 - 4 lambda_k^2] when gamma is not 2 lambda_k. These relations give exact amplitude, variance, and correlation-time targets for optimizer noise.",
 "math_tags": [
  "stochastic-processes",
  "dynamical-systems",
  "probability"
 ],
 "ml_areas": [
  "optimizer",
  "training-dynamics",
  "regularization"
 ],
 "paper": {
  "arxiv_id": "2608.29226",
  "arxiv_url": "https://arxiv.org/abs/2608.29226",
  "summary_what_math_gives_to_ml": "The paper gives an exact construction for bounded, finite-correlation-time nonequilibrium fluctuations: a stable linear relaxation process driven by independent dichotomous telegraph signals. Its transferable assets are a hard stationary support bound, analytically predictable variance and autocorrelation, and a controlled crossover from strongly non-Gaussian behavior at small K to Gaussian behavior at large K. The most direct neural-network transfer is to replace unbounded Gaussian noise in optimizer momentum or latent-state sampling with this bounded colored-noise process. The implementation has sharp falsifiable signatures: the injected state must remain below a known amplitude bound, and its measured moments and correlation curve should match the closed-form predictions.",
  "title": "Ornstein-Uhlenbeck Process Driven by Multiple Dichotomous Noises",
  "year": "2026"
 },
 "ratings": {
  "difficulty": 4,
  "novelty": 7,
  "usefulness": 7
 },
 "solves": [
  "stability",
  "generalization",
  "sample-efficiency"
 ],
 "title": "Bounded Telegraph Exploration for Optimizers",
 "url": "https://synthcore.org/idea/2817/bounded-telegraph-exploration-for-optimizers",
 "verification": {
  "peer_reviewed": false,
  "status": "unverified",
  "status_label": "Unverified",
  "verdict_source": "deterministic test code (paired-seed permutation statistics)"
 }
}