Эксперимент: Persistent Workspace for Online Adaptation (#908)

{ "worked": true, "confidence": 8, "verdict": "Built a gated persistent workspace with explicit write-port/workspace partitioning, shared nonlinear transition, readout, and learned retention gate. The three mechanism predictions were numerically confirmed: retention matched g^n, overwrite matched 1-g, and the stability boundary matched lambda*g <= 1 across the tested sweep. The delayed-bit task showed no practical accuracy win: the ungated baseline reached 1.00 at all gaps, while the gated workspace reached 1.00, 0.495, and 0.529 for gaps 2, 8, and 16; therefore the mechanism manifested, but this setup harmed long-range task performance.", "metrics": { "baseline": "Delayed-bit accuracy: gap 2=1.000, gap 8=1.000, gap 16=1.000; analytic retention and stability predictions matched the tested cases.", "idea": "Gated workspace delayed-bit accuracy: gap 2=1.000, gap 8=0.495, gap 16=0.529; retention at g=0.8,n=20 was observed/predicted 0.011529/0.011529 and at g=0.99,n=20 was 0.817907/0.817907." }, "how_to_run": "/home/maxwelhelp/main/bin/python3 persistent_workspace_experiment.py", "files": [ "persistent_workspace_experiment.py", "results.json" ], "limitations": "Only a tiny synthetic delayed-bit task was tested, not sequential MNIST, language modeling, S4, or distribution-switch adaptation. The baseline was an ungated version of this toy workspace rather than a separately tuned GRU, and only one random seed and training configuration were used." }