Эксперимент: FDT-Calibrated Rotational Optimizer (#555)

{ "worked": true, "confidence": 9, "verdict": "Built the rank-two skew rotational optimizer and numerically verified the core mechanism. The toy sweep confirmed the predicted Euler stability boundary (predicted 1.2, measured 1.2), damped-envelope scaling, oscillation frequency, and Lyapunov covariance to near machine precision. However, the small MLP experiment showed no optimization win: rotational MSE was 0.03645 versus 0.01916 for momentum SGD and 0.02089 for Adam, with higher runtime.", "metrics": { "baseline": "Momentum SGD: final MSE 0.0191641, 0.928 s; Adam: final MSE 0.0208938, 1.239 s; SGD: final MSE 0.0407253, 1.157 s", "idea": "Rotational optimizer: final MSE 0.0364512, 2.695 s; toy boundary a*=1.2 predicted and 1.2 measured; envelope and frequency predictions matched to approximately 1e-15" }, "how_to_run": "/home/maxwelhelp/main/bin/python3 experiment.py", "files": [ "experiment.py", "results.json", "REPORT.md" ], "limitations": "The finite-difference Jacobian/Arnoldi controller, MNIST comparison, empirical fluctuation-response residual, and real-network alpha sweep were not implemented. The neural test used one seed and a small synthetic regression task; the toy covariance check used isotropic noise, so it did not test alpha-dependent FDT residuals." }