Эксперимент: PEP-Synthesized Minimax Optimizer (#855)

{ "worked": false, "confidence": 9, "verdict": "Built a reproducible short-memory minimax recurrence search and compared it with tuned simultaneous GDA on the scalar quadratic game. The searched recurrence reduced worst-case design spectral radius from 0.9978 to 0.9476, but the supplied interpolation matrix produced negative quadratic forms for 40.2% of valid sampled scalar quadratics, so its stated nonnegativity claim failed. The intended mechanism therefore was not validated; additionally, the synthesized rho was essentially flat rather than increasing with coupling, and both methods remained stable through the tested design range.", "metrics": { "baseline": "Tuned simultaneous GDA: worst-case spectral radius 0.997785 on coupling A in [0,1.5]; rho(A=0)=0.995575 and rho(A=1.5)=0.997785.", "idea": "Constrained N=1 memory recurrence coefficients [a1,b0,b1]=[0.665162,0.978414,-0.082054], with a0=1-a1: worst-case spectral radius 0.947600 on A in [0,1.5]; rho(A=0)=0.947600 and rho(A=1.5)=0.947600. Interpolation check: supplied M minimum -4.16683, negative fraction 0.402; claimed nonnegativity not observed." }, "how_to_run": "/home/maxwelhelp/main/bin/python3 pep_minimax.py", "files": [ "pep_minimax.py", "results.json" ], "limitations": "This is a scalar quadratic spectral-radius MVP, not a full interpolation SDP or neural/GAN experiment. No Lyapunov certificate, multidimensional matrix coupling, Hessian estimation, FLOP-matched training, or statistical multi-seed evaluation was performed. The interpolation formula was tested exactly as supplied; its sign/convention issue was not repaired into a new certified method." }