Эксперимент: Autocatalytic Hysteresis Memory Cell (#528)

{ "worked": false, "confidence": 9, "verdict": "Built a readable deterministic projected-Euler autocatalytic cell study with fixed-point, frequency, noise, and toy classification checks. The stability pattern for three real roots and a frequency-area peak near omegatau=0.95 were observed, but the middle root was negative, slow-forcing area remained large instead of vanishing, and no noise turnover appeared; the trivial classifier showed no advantage over the baseline. Therefore the proposed mechanism was not validated as stated.", "metrics": { "baseline": "Linear memory accuracy 1.0; hysteresis comparison not applicable", "idea": "Cell accuracy 1.0; area peak 12.4065 at omegatau=0.9487, low-frequency area 106.06 at omega*tau=0.03, noise area 12.34 (D=0) to 13.07 (D=0.02)" }, "how_to_run": "/home/maxwelhelp/main/bin/python3 experiment.py", "files": [ "experiment.py", "results.json", "report.json" ], "limitations": "Only a scalar deterministic toy cell and a trivial synthetic classification task were tested; no GRU, learned multivariate recurrent model, permuted-MNIST, chemical master-equation simulation, parameter-count matching, GPU training, or broader noise/forcing calibration was performed." }