Uncertainty-guided family sampling / results.json

Mechanism confirmed, baseline not beaten

Raw ⬇ ZIP
  1{
  2  "setup": {
  3    "families": 32,
  4    "budget_including_initial_counts": 4096,
  5    "P_min": 0.0010094396057818674,
  6    "P_max": 0.1498141208173465
  7  },
  8  "prediction_1_variance_identity": {
  9    "predicted": 1.257636789757317,
 10    "observed": 1.2291597616436916,
 11    "relative_error": 0.0226432848860287
 12  },
 13  "prediction_2_optimal_allocation_sweep": [
 14    {
 15      "budget": 512,
 16      "uniform_predicted_variance": 10.061094318058537,
 17      "u_proportional_predicted_variance": 6.883312075843824,
 18      "ratio_uniform_over_u_proportional": 1.4616647054790335
 19    },
 20    {
 21      "budget": 1024,
 22      "uniform_predicted_variance": 5.030547159029268,
 23      "u_proportional_predicted_variance": 3.439300430380218,
 24      "ratio_uniform_over_u_proportional": 1.462665812673171
 25    },
 26    {
 27      "budget": 2048,
 28      "uniform_predicted_variance": 2.515273579514634,
 29      "u_proportional_predicted_variance": 1.7194915385187661,
 30      "ratio_uniform_over_u_proportional": 1.4628007891689796
 31    },
 32    {
 33      "budget": 4096,
 34      "uniform_predicted_variance": 1.257636789757317,
 35      "u_proportional_predicted_variance": 0.859729387222611,
 36      "ratio_uniform_over_u_proportional": 1.4628286626564682
 37    }
 38  ],
 39  "prediction_3_exploration_sweep": [
 40    {
 41      "epsilon": 0.0,
 42      "count_score_correlation": 0.9916290367781933,
 43      "minimum_family_share": 0.0078125,
 44      "predicted_variance": 0.9029224084969325
 45    },
 46    {
 47      "epsilon": 0.01,
 48      "count_score_correlation": 0.991826823932279,
 49      "minimum_family_share": 0.008056640625,
 50      "predicted_variance": 0.9009406137305294
 51    },
 52    {
 53      "epsilon": 0.05,
 54      "count_score_correlation": 0.992740599014048,
 55      "minimum_family_share": 0.009033203125,
 56      "predicted_variance": 0.8986331886186223
 57    },
 58    {
 59      "epsilon": 0.2,
 60      "count_score_correlation": 0.9904190646879453,
 61      "minimum_family_share": 0.011474609375,
 62      "predicted_variance": 0.8982821453574045
 63    },
 64    {
 65      "epsilon": 0.5,
 66      "count_score_correlation": 0.9867226860128363,
 67      "minimum_family_share": 0.016357421875,
 68      "predicted_variance": 0.9224689978774235
 69    }
 70  ],
 71  "prediction_3_mode_collapse_sweep": [
 72    {
 73      "epsilon": 0.0,
 74      "hidden_family": 1,
 75      "hidden_family_count": 3,
 76      "true_predicted_variance": 8.966861069114389
 77    },
 78    {
 79      "epsilon": 0.01,
 80      "hidden_family": 1,
 81      "hidden_family_count": 5,
 82      "true_predicted_variance": 5.6810797215410656
 83    },
 84    {
 85      "epsilon": 0.05,
 86      "hidden_family": 1,
 87      "hidden_family_count": 10,
 88      "true_predicted_variance": 3.2191381989676686
 89    },
 90    {
 91      "epsilon": 0.2,
 92      "hidden_family": 1,
 93      "hidden_family_count": 33,
 94      "true_predicted_variance": 1.5058666088665276
 95    }
 96  ],
 97  "equal_budget_comparison": {
 98    "uniform": {
 99      "predicted_variance": 1.257636789757317,
100      "empirical_variance": 1.2911391289589171
101    },
102    "magnitude_only": {
103      "predicted_variance": 0.912940780116989,
104      "empirical_variance": 0.9170077821485625
105    },
106    "uncertainty_guided_controller": {
107      "predicted_variance": 0.8986331886186223,
108      "empirical_variance": 0.8928061805630432
109    }
110  }
111}