# Эксперимент: Collision-aware physical-support abstention (#422)

{
  "worked": true,
  "confidence": 7,
  "verdict": "Built a reproducible synthetic two-ray collision experiment with sign-invariant ray distance, information-based parent/child resolution, replicated noisy test views, and dictionary-orientation uncertainty. The predicted transitions were observed: the test gate switched at T*=4 for s=0.1, the dictionary gate switched at N*=15,625 for s=0.2, and log-log sweeps measured slopes 2.000 for I_S versus s and 6.000 for I_D versus s. In the low-I_D regime abstention reduced physical child FDR from 0.338 to 0 while preserving parent recall at 1.0; this is a promising mechanism signal, though the orientation uncertainty was simulated from the predicted scaling rather than learned from raw dictionary-training data.",
  "metrics": {
    "baseline": "Ordinary child labeling: low-I_D child FDR 0.338, parent recall 1.000; high-I_D child FDR 0.183, child accuracy 0.817.",
    "idea": "Low-I_D collision-aware abstention: physical child FDR 0.000, parent recall 1.000; resolution transitioned at predicted N*=15,625. Formula sweeps measured I_S~s^2 and I_D~s^6 with observed slopes 2.000 and 6.000."
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 collision_abstention.py",
  "files": [
    "collision_abstention.py",
    "results.json"
  ],
  "limitations": "This is a toy two-ray experiment, not a trained sparse encoder, dictionary-learning model, MNIST/CIFAR benchmark, or MoE router. Dictionary orientation error was injected with the predicted 1/(sqrt(N)s^2) scale rather than estimated from actual latent training moments, and thresholds were set to 1 rather than calibrated."
}