Эксперимент: PSD Spectral CNN Block (#1173)

{ "worked": true, "confidence": 8, "verdict": "Built a circular multi-channel PSD convolution block as BB, with an explicit adjoint and Fourier-spectrum checker. The core claim was numerically confirmed: adjoint relative error was 4.7e-8, the minimum sampled Fourier eigenvalue was 0.198, and the diffusion step using tau=0.1383 had maximum tested norm ratio 0.788. The tiny denoising comparison did not show an accuracy win: PSD test MSE was 0.10220 versus 0.10209 for the unconstrained baseline, while using 54 versus 36 parameters.", "metrics": { "baseline": "Unconstrained residual convolution: test MSE 0.1020901, train MSE 0.1087475, 36 parameters", "idea": "PSD BB residual block: test MSE 0.1021957, train MSE 0.1086864, 54 parameters; Fourier minimum eigenvalue 0.198356, maximum 7.230365, diffusion norm ratio 0.787504" }, "how_to_run": "/home/maxwelhelp/main/bin/python3 run_experiment.py", "files": [ "run_experiment.py", "results.json" ], "limitations": "This was a small synthetic denoising experiment, not CIFAR-10 or a six-layer CNN. Only sampled Fourier frequencies and random-vector diffusion ratios were checked; no exhaustive operator-norm computation, frequency-grid penalty baseline, robustness sweep, gradient-spike analysis, FLOP-normalized comparison, or multiple-seed statistical evaluation was performed." }