Monotone Compositional Reachability Critic / results.json

Mechanism confirmed, baseline not beaten

Raw ⬇ ZIP
  1{
  2  "predictions": {
  3    "P1_monotone_derivatives_nonnegative": {
  4      "predicted": "100% samples",
  5      "observed_fraction": 1.0,
  6      "tolerance": "exact up to numerical precision"
  7    },
  8    "P2_error_bound": {
  9      "predicted": "|delta V| <= 0.7|e1|+1.4|e2| and linear scaling",
 10      "max_ratio_observed": 1.0007299184799194,
 11      "loglog_scaling_exponent": 0.999923480526178,
 12      "expected_exponent": 1.0
 13    },
 14    "P3_negative_crossing": {
 15      "predicted": "set/order failure begins at c2<0",
 16      "sweep": [
 17        {
 18          "c2": 1.0,
 19          "derivative": 1.0,
 20          "set_disagreement": 0.02319999970495701,
 21          "order_reversal_fraction": 0.0
 22        },
 23        {
 24          "c2": 0.5,
 25          "derivative": 0.5,
 26          "set_disagreement": 0.041999999433755875,
 27          "order_reversal_fraction": 0.0
 28        },
 29        {
 30          "c2": 0.2,
 31          "derivative": 0.2,
 32          "set_disagreement": 0.11640000343322754,
 33          "order_reversal_fraction": 0.0
 34        },
 35        {
 36          "c2": 0.05,
 37          "derivative": 0.05,
 38          "set_disagreement": 0.18039999902248383,
 39          "order_reversal_fraction": 0.0
 40        },
 41        {
 42          "c2": 0.0,
 43          "derivative": 0.0,
 44          "set_disagreement": 0.20960000157356262,
 45          "order_reversal_fraction": 0.0
 46        },
 47        {
 48          "c2": -0.02,
 49          "derivative": -0.02,
 50          "set_disagreement": 0.2231999933719635,
 51          "order_reversal_fraction": 1.0
 52        },
 53        {
 54          "c2": -0.05,
 55          "derivative": -0.05,
 56          "set_disagreement": 0.24040000140666962,
 57          "order_reversal_fraction": 1.0
 58        },
 59        {
 60          "c2": -0.1,
 61          "derivative": -0.1,
 62          "set_disagreement": 0.2678000032901764,
 63          "order_reversal_fraction": 1.0
 64        },
 65        {
 66          "c2": -0.2,
 67          "derivative": -0.2,
 68          "set_disagreement": 0.32339999079704285,
 69          "order_reversal_fraction": 1.0
 70        },
 71        {
 72          "c2": -0.5,
 73          "derivative": -0.5,
 74          "set_disagreement": 0.40540000796318054,
 75          "order_reversal_fraction": 1.0
 76        },
 77        {
 78          "c2": -1.0,
 79          "derivative": -1.0,
 80          "set_disagreement": 0.49799999594688416,
 81          "order_reversal_fraction": 1.0
 82        }
 83      ]
 84    }
 85  },
 86  "mini_experiment": {
 87    "positive_aggregator_weights": [
 88      0.9989890456199646,
 89      0.7955543994903564
 90    ],
 91    "unconstrained_weights": [
 92      0.9989889860153198,
 93      0.7955544590950012
 94    ],
 95    "heldout_rmse": {
 96      "monotone": 0.031863730400800705,
 97      "unconstrained": 0.031863726675510406
 98    },
 99    "heldout_set_iou": {
100      "monotone": 0.9279778599739075,
101      "unconstrained": 0.9279778599739075
102    }
103  }
104}