Hankel Residual Observer / bench_report.json

Failed on benchmark

Raw ⬇ ZIP
  1{
  2  "bench_version": 1,
  3  "track": "sequence",
  4  "model": "transformer_tiny",
  5  "metric_direction": "lower is better",
  6  "n_seeds": 8,
  7  "baseline": {
  8    "best_cfg": {
  9      "lr": 0.001
 10    },
 11    "sweep": [
 12      {
 13        "cfg": {
 14          "lr": 0.001
 15        },
 16        "mean": 0.5443045374850393
 17      },
 18      {
 19        "cfg": {
 20          "lr": 0.003
 21        },
 22        "mean": 0.6654394097127154
 23      },
 24      {
 25        "cfg": {
 26          "lr": 0.006
 27        },
 28        "mean": 0.8961613556937273
 29      }
 30    ],
 31    "full": {
 32      "mean": 0.5443045374850393,
 33      "std": 0.034391634415997095,
 34      "per_seed": [
 35        0.563120002561451,
 36        0.4667357694115086,
 37        0.5482129393734809,
 38        0.5391273206172196,
 39        0.5435938104924675,
 40        0.542868393619944,
 41        0.5989920214133407,
 42        0.5517860423909019
 43      ],
 44      "n": 8
 45    }
 46  },
 47  "idea": {
 48    "mean": 0.7173206874260685,
 49    "std": 0.04324948219202462,
 50    "per_seed": [
 51      0.7285622398292526,
 52      0.6195935527509657,
 53      0.7682973969911581,
 54      0.7145800123187942,
 55      0.7425737796077112,
 56      0.6909476920862011,
 57      0.7539289132268427,
 58      0.7200819125976213
 59    ],
 60    "n": 8
 61  },
 62  "comparison": {
 63    "delta_mean": 0.1730161499410291,
 64    "idea_wins": 0,
 65    "n_pairs": 8,
 66    "per_seed_diffs": [
 67      0.1654422372678016,
 68      0.1528577833394571,
 69      0.22008445761767714,
 70      0.17545269170157463,
 71      0.19897996911524374,
 72      0.14807929846625711,
 73      0.154936891813502,
 74      0.16829587020671943
 75    ],
 76    "p_value": 0.0081,
 77    "mde": 0.020796525052684568,
 78    "mde_rel_pct": 3.8207517337215253,
 79    "verdict": "idea worse (significant)",
 80    "system_worked": false
 81  },
 82  "idea_sweep": [
 83    {
 84      "cfg": {
 85        "lr": 0.001
 86      },
 87      "mean": 0.7173206874260685,
 88      "std": 0.04324948219202462
 89    },
 90    {
 91      "cfg": {
 92        "lr": 0.003
 93      },
 94      "mean": 1.0912813241829704,
 95      "std": 0.5463389443333877
 96    },
 97    {
 98      "cfg": {
 99        "lr": 0.006
100      },
101      "mean": 1.320354687932311,
102      "std": 0.6682789001876265
103    }
104  ],
105  "protocol": {
106    "seeds": [
107      0,
108      1,
109      2,
110      3,
111      4,
112      5,
113      6,
114      7
115    ],
116    "epochs": 8,
117    "n_train": 400,
118    "n_test": 400,
119    "shared_lr_grid": [
120      {
121        "lr": 0.001
122      },
123      {
124        "lr": 0.003
125      },
126      {
127        "lr": 0.006
128      }
129    ],
130    "observer": {
131      "t_ini": 8,
132      "horizon": 1,
133      "max_cols": 40,
134      "ridge": 0.01,
135      "tau": 1.0
136    }
137  },
138  "mechanism_signature": {
139    "residual_forecast_mse": 0.7368731066703664,
140    "residual_naive_mse": 0.545037575640239,
141    "correction_mse": 0.7173206874260685,
142    "base_mse": 0.5443045374850393,
143    "mean_gamma": 0.9632895999502775,
144    "active_fraction": 0.9775,
145    "mean_mismatch_q": 0.01453749365700519,
146    "confirmed": false
147  }
148}