# Стенд-проверка (stage-2) · промт оператора:

(универсальный)

Ответ агента:

{
  "worked": false,
  "confidence": 9,
  "verdict": "On the registered geometric_graph_diffusion benchmark, teleporting diffusion at alpha=0.1 reduced mean MSE from approximately 0.01771914 to 0.01686609, but the paired permutation test gave p=0.14395, so the improvement is not significant. The mechanism signature was also not confirmed: trained-model contraction differed from the nominal 1-alpha prediction by up to 0.39057. Therefore the system correctly reports no significant win.",
  "metrics": {
    "baseline": "Registered track geometric_graph_diffusion; swept alpha=0, lr in {0.003,0.01,0.03}, weight decay in {0,0.0001}, 20 epochs; selected lr=0.03, wd=0.0001. Full 8-seed mean MSE approximately 0.01771914.",
    "idea": "Registered track, same architecture and selected training hyperparameters, alpha sweep {0.03,0.1,0.3}; best alpha=0.1, mean MSE 0.01686609, paired delta -0.00085305, 6/8 wins, p=0.14395."
  },
  "bench_report": {
    "bench_version": 1,
    "track": "geometric_graph_diffusion",
    "model": "graph_diffusion_net",
    "metric_direction": "lower is better",
    "n_seeds": 8,
    "baseline": {
      "best_cfg": {
        "alpha": 0.0,
        "lr": 0.03,
        "wd": 0.0001,
        "epochs": 20
      },
      "sweep": "Complete baseline sweep is stored in bench_report.json; it covers all six shared lr/weight-decay configurations.",
      "full": {
        "mean": 0.017719140276312828,
        "per_seed": "Stored in bench_report.json",
        "n": 8
      }
    },
    "idea": {
      "mean": 0.016866090474650264,
      "std": 0.009188740388655336,
      "per_seed": [
        0.013914111070334911,
        0.01861773431301117,
        0.013892930932343006,
        0.010849558748304844,
        0.03997589275240898,
        0.014381146989762783,
        0.008187311701476574,
        0.015110037289559841
      ],
      "n": 8
    },
    "comparison": {
      "delta_mean": -0.0008530498016625643,
      "idea_wins": 6,
      "n_pairs": 8,
      "per_seed_diffs": [
        -0.0016245488077402115,
        0.001920647919178009,
        -0.0009526042267680168,
        -0.0012918869033455849,
        -0.002196080982685089,
        0.0008739698678255081,
        -0.002113870345056057,
        -0.0014400249347090721
      ],
      "p_value": 0.14395,
      "verdict": "no significant win",
      "system_worked": false
    },
    "mechanism_signature": {
      "prediction": "uniform teleportation contracts mean-zero feature modes approximately by 1-alpha",
      "trained_model_rows": [
        {
          "alpha": 0.0,
          "observed_mean_centered_norm_ratio": 0.6094285285591718,
          "predicted": 1.0
        },
        {
          "alpha": 0.03,
          "observed_mean_centered_norm_ratio": 0.5925245406162934,
          "predicted": 0.97
        },
        {
          "alpha": 0.1,
          "observed_mean_centered_norm_ratio": 0.5551252809546359,
          "predicted": 0.9
        },
        {
          "alpha": 0.3,
          "observed_mean_centered_norm_ratio": 0.4483752569702065,
          "predicted": 0.7
        }
      ],
      "max_abs_error": 0.39057147144082816,
      "confirmed": false
    },
    "custom_track": {
      "name": "geometric_graph_diffusion",
      "file": "registered_bench.py",
      "domain": "graph-neural-networks"
    }
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 registered_bench.py",
  "files": [
    "registered_bench.py",
    "bench_report.json"
  ],
  "limitations": "The registered benchmark uses a synthetic geometric kNN graph track rather than a real-world simplicial-complex dataset. The intervention uses node-level weighted graph diffusion reconstructed from the registered track; explicit higher-order triangle incidence was not available in that track. The complete baseline sweep and exact baseline per-seed values are retained in bench_report.json.",
  "system_verdict": "failed",
  "practical_verdict": "inconclusive",
  "mechanism_ok": 0,
  "system_judged": true
}