# Эксперимент: Dissipative drift parameterization (#1390)

{
  "worked": true,
  "confidence": 8,
  "verdict": "Built a conditional dissipative drift module b=-mu*a+Lf*tanh(a+h(t,x)) and compared it with an unconstrained MLP on a fixed-seed toy drift-fitting task. The mathematical claim was reproduced: residual finite-difference action Lipschitz was 0.347 < Lf=0.35, the observed one-sided ratio was -0.453 versus the bound -0.45, and sampled violations were zero. Euler rollouts had lower action norms and variance for the dissipative model at 20 and 50 steps, but its validation drift MSE was worse, so the result demonstrates stability rather than an accuracy win.",
  "metrics": {
    "baseline": "Generic MLP: validation drift MSE 0.000833; rollout mean norm/std 1.9549/0.9943 at 8 steps, 1.0475/0.5182 at 20, 0.7777/0.3100 at 50. One-sided violation fraction 1e-4.",
    "idea": "Dissipative drift: validation drift MSE 0.026612; rollout mean norm/std 1.9208/1.0876 at 8 steps, 0.9543/0.4795 at 20, 0.5436/0.1771 at 50. Residual FD Lipschitz 0.3471 <= 0.35, worst one-sided ratio -0.4531 <= -0.45, violation fraction 0."
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 experiment.py",
  "files": [
    "experiment.py",
    "results.json"
  ],
  "limitations": "This is a small synthetic drift-fitting proxy, not a full diffusion-policy or RL benchmark. It uses one random seed, does not test sample-size scaling or the paper's finite-sample rates, and rollout norms are a stability proxy rather than held-out likelihood, mode coverage, or return."
}