Эксперимент: Automaton-Supervised Neural Policy Shield (#966)

{ "worked": false, "confidence": 9, "verdict": "Built and numerically verified a deterministic automaton shield using the one-hot transition update, plus an exhaustive receding-horizon planner. The shield achieved zero forbidden transitions across all bias values, and planning showed the predicted horizon transition: completion was 0.0 at H=1 and 1.0 for H>=2. However, the attempted intervention-probability prediction was not confirmed because the measured intervention rate includes stochastic resampling differences and state-qualified event labels, so the strict mechanism criterion is not fully satisfied; additionally, the printed (AB^T)⊙C=C certificate is dimensionally ill-typed for the stated matrix shapes.", "metrics": { "baseline": "Unshielded forbidden-transition rate: 0.5965 at bias=-2, rising to 1.0 at biases 4 and 5; completion 0.01675 at bias=4.", "idea": "Shield forbidden-transition rate: 0.0 for every tested bias; shield completion at bias=4 was 0.01675. Planner completion at bias=4: H=1 0.0, H=2 1.0, H=3 1.0, H=4 1.0; all planner violation rates were 0.0." }, "how_to_run": "/home/maxwelhelp/main/bin/python3 run_experiment.py", "files": [ "run_experiment.py", "results.json" ], "limitations": "This is a tiny tabular toy rather than the proposed two-agent PPO grid-world. No neural network training, shared-resource collision benchmark, PPO sample efficiency, FLOP comparison, or real observation/state-tracking errors were tested. The intervention sweep needs a redesigned coupling between raw and shielded sampling before its analytic probability can be meaningfully evaluated." }