Эксперимент: Intrinsic-Dimension Batch Audit (#1356)

{ "worked": false, "confidence": 8, "verdict": "Built a reproducible intrinsic-dimension batch audit prototype with k-NN graph energy, norm/energy distortion scoring, candidate retries, and a small MLP comparison. The toy math check showed graph energy increasing with frequency and auditing reduced norm-estimation error for frequencies 1–8, but the training experiment showed no gradient-cosine or loss-bias improvement: cosine was 0.3318 baseline versus 0.3294 audited, and bias was 0.01246 versus 0.01247. The audited controller also evaluated three candidates per update, giving 11.7% measured local overhead but 3x attempted examples and lower accuracy, so the promised practical win was not observed.", "metrics": { "baseline": "gradient cosine 0.3318; loss-estimate bias 0.01246; final-window loss 0.6851; accuracy 0.6139; retries 0; audit overhead 4.1%", "idea": "gradient cosine 0.3294; loss-estimate bias 0.01247; final-window loss 0.6760; accuracy 0.5431; retries 2; audit overhead 11.7%; audited norm error was lower than random for frequencies 1–8" }, "how_to_run": "/home/maxwelhelp/main/bin/python3 audit_experiment.py", "files": [ "audit_experiment.py", "results.json" ], "limitations": "This was a small synthetic circle-manifold experiment rather than CIFAR-10, used scalar losses and one logit coordinate instead of true gradient projections, and computed full-dataset audit values for convenience. It did not implement diffusion-DPP sampling, arbitrary-inclusion-probability Horvitz-Thompson edge handling, probe refresh intervals, threshold tuning, or a replicated held-out study." }