Эксперимент: Heisenberg latent upsampler (#1300)

{ "worked": false, "confidence": 9, "verdict": "Built the Heisenberg refinement module and numerically verified the core group behavior: associativity error was 1.78e-15, and reversing a unit-square path changed area from +1 to -1 exactly. However, on 100 fixed-seed smooth trajectories, the proposed correction had central-coordinate RMSE 0.02558 versus 0.000328 for ordinary four-point interpolation and 0.001315 for linear interpolation, so the claimed interpolation benefit was not observed in this toy test.", "metrics": { "baseline": "Ordinary four-point z RMSE mean 0.0003284 (median 0.0003008); linear RMSE mean 0.0013154.", "idea": "Heisenberg four-point refinement z RMSE mean 0.02558 (median 0.02319); mean max |z| 2.07284 versus 2.07185 for ordinary four-point." }, "how_to_run": "/home/maxwelhelp/main/bin/python3 heisenberg_experiment.py", "files": [ "heisenberg_experiment.py", "results.json" ], "limitations": "Only a synthetic interpolation test was run; no learned RNN/world-model task, FLOP-matched training comparison, gradient study, or repeated-refinement stability sweep was performed. The negative result may reflect that the correction preserves path geometry rather than reconstructing the particular fine-grid cumulative-area coordinate used as the RMSE target." }