Add severity distribution comparison results

severity_distribution_comparison.json

@@ -35,8 +35,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "4014": {
           "n": 30,
@@ -47,8 +47,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "F02": {
           "n": 200,
@@ -124,21 +124,21 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.0,
-        "Moderate": 0.5,
+        "Mild": 0.0875,
+        "Moderate": 0.25,
         "Normal": 0.325,
-        "Severe": 0.175
+        "Severe": 0.3375
       },
       "ground_truth_counts": {
-        "Mild": 0,
-        "Moderate": 400,
+        "Mild": 70,
+        "Moderate": 200,
         "Normal": 260,
-        "Severe": 140
+        "Severe": 270
       },
       "gt_valid_samples": 800,
       "gt_unknown_samples": 0,
-      "classifier_accuracy_on_dataset": 0.6987,
-      "speaker_severity_agreement": 0.6667,
+      "classifier_accuracy_on_dataset": 0.5675,
+      "speaker_severity_agreement": 0.75,
       "distance_metrics": {
         "jensen_shannon": 0.0,
         "cosine_similarity": 1.0,
@@ -180,8 +180,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "4014": {
           "n": 30,
@@ -192,8 +192,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "F02": {
           "n": 200,
@@ -269,21 +269,21 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.094595,
-        "Moderate": 0.27027,
-        "Normal": 0.27027,
-        "Severe": 0.364865
+        "Mild": 0.0875,
+        "Moderate": 0.25,
+        "Normal": 0.325,
+        "Severe": 0.3375
       },
       "ground_truth_counts": {
         "Mild": 70,
         "Moderate": 200,
-        "Normal": 200,
+        "Normal": 260,
         "Severe": 270
       },
-      "gt_valid_samples": 740,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.3014,
-      "speaker_severity_agreement": 0.3333,
+      "gt_valid_samples": 800,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.3538,
+      "speaker_severity_agreement": 0.5,
       "distance_metrics": {
         "jensen_shannon": 0.431928,
         "cosine_similarity": 0.749113,
@@ -291,10 +291,10 @@
         "total_variation": 0.49
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.26171,
-        "cosine_similarity": 0.859725,
-        "kl_divergence": 0.238945,
-        "total_variation": 0.28446
+        "jensen_shannon": 0.0,
+        "cosine_similarity": 1.0,
+        "kl_divergence": 0.0,
+        "total_variation": 0.0
       }
     },
     "orpheus_tts": {
@@ -414,20 +414,20 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.098712,
-        "Moderate": 0.260372,
-        "Normal": 0.27897,
-        "Severe": 0.361946
+        "Mild": 0.090909,
+        "Moderate": 0.239789,
+        "Normal": 0.335968,
+        "Severe": 0.333333
       },
       "ground_truth_counts": {
         "Mild": 69,
         "Moderate": 182,
-        "Normal": 195,
+        "Normal": 255,
         "Severe": 253
       },
-      "gt_valid_samples": 699,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.4263,
+      "gt_valid_samples": 759,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.4717,
       "speaker_severity_agreement": 0.6,
       "distance_metrics": {
         "jensen_shannon": 0.041223,
@@ -436,10 +436,10 @@
         "total_variation": 0.042764
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.266293,
-        "cosine_similarity": 0.854508,
-        "kl_divergence": 0.248709,
-        "total_variation": 0.285658
+        "jensen_shannon": 0.011022,
+        "cosine_similarity": 0.999562,
+        "kl_divergence": 0.000486,
+        "total_variation": 0.014378
       }
     },
     "sesame_tts": {
@@ -559,20 +559,20 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.098712,
-        "Moderate": 0.260372,
-        "Normal": 0.27897,
-        "Severe": 0.361946
+        "Mild": 0.090909,
+        "Moderate": 0.239789,
+        "Normal": 0.335968,
+        "Severe": 0.333333
       },
       "ground_truth_counts": {
         "Mild": 69,
         "Moderate": 182,
-        "Normal": 195,
+        "Normal": 255,
         "Severe": 253
       },
-      "gt_valid_samples": 699,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.4621,
+      "gt_valid_samples": 759,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.5046,
       "speaker_severity_agreement": 0.6,
       "distance_metrics": {
         "jensen_shannon": 0.062168,
@@ -581,10 +581,10 @@
         "total_variation": 0.071675
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.266293,
-        "cosine_similarity": 0.854508,
-        "kl_divergence": 0.248709,
-        "total_variation": 0.285658
+        "jensen_shannon": 0.011022,
+        "cosine_similarity": 0.999562,
+        "kl_divergence": 0.000486,
+        "total_variation": 0.014378
       }
     },
     "spark_tts": {
@@ -704,20 +704,20 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.093793,
-        "Moderate": 0.270345,
-        "Normal": 0.273103,
-        "Severe": 0.362759
+        "Mild": 0.086624,
+        "Moderate": 0.249682,
+        "Normal": 0.328662,
+        "Severe": 0.335032
       },
       "ground_truth_counts": {
         "Mild": 68,
         "Moderate": 196,
-        "Normal": 198,
+        "Normal": 258,
         "Severe": 263
       },
-      "gt_valid_samples": 725,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.5586,
+      "gt_valid_samples": 785,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.5924,
       "speaker_severity_agreement": 0.25,
       "distance_metrics": {
         "jensen_shannon": 0.051367,
@@ -726,10 +726,10 @@
         "total_variation": 0.042508
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.260292,
-        "cosine_similarity": 0.861648,
-        "kl_divergence": 0.23643,
-        "total_variation": 0.281552
+        "jensen_shannon": 0.002923,
+        "cosine_similarity": 0.999965,
+        "kl_divergence": 3.4e-05,
+        "total_variation": 0.003662
       }
     },
     "knn_vc": {
@@ -849,20 +849,20 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.094723,
-        "Moderate": 0.270636,
-        "Normal": 0.269283,
-        "Severe": 0.365359
+        "Mild": 0.08761,
+        "Moderate": 0.250313,
+        "Normal": 0.324155,
+        "Severe": 0.337922
       },
       "ground_truth_counts": {
         "Mild": 70,
         "Moderate": 200,
-        "Normal": 199,
+        "Normal": 259,
         "Severe": 270
       },
-      "gt_valid_samples": 739,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.5007,
+      "gt_valid_samples": 799,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.5382,
       "speaker_severity_agreement": 0.6,
       "distance_metrics": {
         "jensen_shannon": 0.390763,
@@ -871,10 +871,10 @@
         "total_variation": 0.457436
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.261905,
-        "cosine_similarity": 0.859496,
-        "kl_divergence": 0.239221,
-        "total_variation": 0.285081
+        "jensen_shannon": 0.000638,
+        "cosine_similarity": 0.999998,
+        "kl_divergence": 2e-06,
+        "total_variation": 0.000845
       }
     },
     "orpheus_knn_vc": {
@@ -994,20 +994,20 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.098712,
-        "Moderate": 0.260372,
-        "Normal": 0.27897,
-        "Severe": 0.361946
+        "Mild": 0.090909,
+        "Moderate": 0.239789,
+        "Normal": 0.335968,
+        "Severe": 0.333333
       },
       "ground_truth_counts": {
         "Mild": 69,
         "Moderate": 182,
-        "Normal": 195,
+        "Normal": 255,
         "Severe": 253
       },
-      "gt_valid_samples": 699,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.6853,
+      "gt_valid_samples": 759,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.7088,
       "speaker_severity_agreement": 0.5,
       "distance_metrics": {
         "jensen_shannon": 0.295507,
@@ -1016,10 +1016,10 @@
         "total_variation": 0.308467
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.266293,
-        "cosine_similarity": 0.854508,
-        "kl_divergence": 0.248709,
-        "total_variation": 0.285658
+        "jensen_shannon": 0.011022,
+        "cosine_similarity": 0.999562,
+        "kl_divergence": 0.000486,
+        "total_variation": 0.014378
       }
     },
     "sesame_knn_vc": {
@@ -1050,8 +1050,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "4014": {
           "n": 30,
@@ -1062,8 +1062,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "F02": {
           "n": 182,
@@ -1139,21 +1139,21 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.098712,
-        "Moderate": 0.260372,
-        "Normal": 0.27897,
-        "Severe": 0.361946
+        "Mild": 0.090909,
+        "Moderate": 0.239789,
+        "Normal": 0.335968,
+        "Severe": 0.333333
       },
       "ground_truth_counts": {
         "Mild": 69,
         "Moderate": 182,
-        "Normal": 195,
+        "Normal": 255,
         "Severe": 253
       },
-      "gt_valid_samples": 699,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.6924,
-      "speaker_severity_agreement": 0.8333,
+      "gt_valid_samples": 759,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.7141,
+      "speaker_severity_agreement": 0.875,
       "distance_metrics": {
         "jensen_shannon": 0.230947,
         "cosine_similarity": 0.858436,
@@ -1161,10 +1161,10 @@
         "total_variation": 0.231335
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.266293,
-        "cosine_similarity": 0.854508,
-        "kl_divergence": 0.248709,
-        "total_variation": 0.285658
+        "jensen_shannon": 0.011022,
+        "cosine_similarity": 0.999562,
+        "kl_divergence": 0.000486,
+        "total_variation": 0.014378
       }
     },
     "spark_knn_vc": {
@@ -1195,8 +1195,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "4014": {
           "n": 30,
@@ -1207,8 +1207,8 @@
             "Severe": 0.0
           },
           "majority_class": "Normal",
-          "expected_severity": "Unknown",
-          "severity_match": null
+          "expected_severity": "Normal",
+          "severity_match": true
         },
         "F02": {
           "n": 196,
@@ -1284,21 +1284,21 @@
         }
       },
       "ground_truth_distribution": {
-        "Mild": 0.093793,
-        "Moderate": 0.270345,
-        "Normal": 0.273103,
-        "Severe": 0.362759
+        "Mild": 0.086624,
+        "Moderate": 0.249682,
+        "Normal": 0.328662,
+        "Severe": 0.335032
       },
       "ground_truth_counts": {
         "Mild": 68,
         "Moderate": 196,
-        "Normal": 198,
+        "Normal": 258,
         "Severe": 263
       },
-      "gt_valid_samples": 725,
-      "gt_unknown_samples": 60,
-      "classifier_accuracy_on_dataset": 0.6924,
-      "speaker_severity_agreement": 0.8333,
+      "gt_valid_samples": 785,
+      "gt_unknown_samples": 0,
+      "classifier_accuracy_on_dataset": 0.7159,
+      "speaker_severity_agreement": 0.875,
       "distance_metrics": {
         "jensen_shannon": 0.258646,
         "cosine_similarity": 0.823941,
@@ -1306,10 +1306,10 @@
         "total_variation": 0.254961
       },
       "gt_distance_metrics": {
-        "jensen_shannon": 0.260292,
-        "cosine_similarity": 0.861648,
-        "kl_divergence": 0.23643,
-        "total_variation": 0.281552
+        "jensen_shannon": 0.002923,
+        "cosine_similarity": 0.999965,
+        "kl_divergence": 3.4e-05,
+        "total_variation": 0.003662
       }
     }
   },
@@ -1395,95 +1395,95 @@
       "kl_divergence": 0.0,
       "total_variation": 0.0,
       "n": 800,
-      "accuracy": 0.6987,
-      "speaker_agreement": 0.6667
+      "accuracy": 0.5675,
+      "speaker_agreement": 0.75
     },
     {
-      "name": "spark_tts",
-      "jensen_shannon": 0.260292,
-      "cosine_similarity": 0.861648,
-      "kl_divergence": 0.23643,
-      "total_variation": 0.281552,
-      "n": 725,
-      "accuracy": 0.5586,
-      "speaker_agreement": 0.25
+      "name": "baseline_orpheus",
+      "jensen_shannon": 0.0,
+      "cosine_similarity": 1.0,
+      "kl_divergence": 0.0,
+      "total_variation": 0.0,
+      "n": 800,
+      "accuracy": 0.3538,
+      "speaker_agreement": 0.5
     },
     {
-      "name": "spark_knn_vc",
-      "jensen_shannon": 0.260292,
-      "cosine_similarity": 0.861648,
-      "kl_divergence": 0.23643,
-      "total_variation": 0.281552,
-      "n": 725,
-      "accuracy": 0.6924,
-      "speaker_agreement": 0.8333
+      "name": "knn_vc",
+      "jensen_shannon": 0.000638,
+      "cosine_similarity": 0.999998,
+      "kl_divergence": 2e-06,
+      "total_variation": 0.000845,
+      "n": 799,
+      "accuracy": 0.5382,
+      "speaker_agreement": 0.6
     },
     {
-      "name": "baseline_orpheus",
-      "jensen_shannon": 0.26171,
-      "cosine_similarity": 0.859725,
-      "kl_divergence": 0.238945,
-      "total_variation": 0.28446,
-      "n": 740,
-      "accuracy": 0.3014,
-      "speaker_agreement": 0.3333
+      "name": "spark_tts",
+      "jensen_shannon": 0.002923,
+      "cosine_similarity": 0.999965,
+      "kl_divergence": 3.4e-05,
+      "total_variation": 0.003662,
+      "n": 785,
+      "accuracy": 0.5924,
+      "speaker_agreement": 0.25
     },
     {
-      "name": "knn_vc",
-      "jensen_shannon": 0.261905,
-      "cosine_similarity": 0.859496,
-      "kl_divergence": 0.239221,
-      "total_variation": 0.285081,
-      "n": 739,
-      "accuracy": 0.5007,
-      "speaker_agreement": 0.6
+      "name": "spark_knn_vc",
+      "jensen_shannon": 0.002923,
+      "cosine_similarity": 0.999965,
+      "kl_divergence": 3.4e-05,
+      "total_variation": 0.003662,
+      "n": 785,
+      "accuracy": 0.7159,
+      "speaker_agreement": 0.875
     },
     {
       "name": "orpheus_tts",
-      "jensen_shannon": 0.266293,
-      "cosine_similarity": 0.854508,
-      "kl_divergence": 0.248709,
-      "total_variation": 0.285658,
-      "n": 699,
-      "accuracy": 0.4263,
+      "jensen_shannon": 0.011022,
+      "cosine_similarity": 0.999562,
+      "kl_divergence": 0.000486,
+      "total_variation": 0.014378,
+      "n": 759,
+      "accuracy": 0.4717,
       "speaker_agreement": 0.6
     },
     {
       "name": "sesame_tts",
-      "jensen_shannon": 0.266293,
-      "cosine_similarity": 0.854508,
-      "kl_divergence": 0.248709,
-      "total_variation": 0.285658,
-      "n": 699,
-      "accuracy": 0.4621,
+      "jensen_shannon": 0.011022,
+      "cosine_similarity": 0.999562,
+      "kl_divergence": 0.000486,
+      "total_variation": 0.014378,
+      "n": 759,
+      "accuracy": 0.5046,
       "speaker_agreement": 0.6
     },
     {
       "name": "orpheus_knn_vc",
-      "jensen_shannon": 0.266293,
-      "cosine_similarity": 0.854508,
-      "kl_divergence": 0.248709,
-      "total_variation": 0.285658,
-      "n": 699,
-      "accuracy": 0.6853,
+      "jensen_shannon": 0.011022,
+      "cosine_similarity": 0.999562,
+      "kl_divergence": 0.000486,
+      "total_variation": 0.014378,
+      "n": 759,
+      "accuracy": 0.7088,
       "speaker_agreement": 0.5
     },
     {
       "name": "sesame_knn_vc",
-      "jensen_shannon": 0.266293,
-      "cosine_similarity": 0.854508,
-      "kl_divergence": 0.248709,
-      "total_variation": 0.285658,
-      "n": 699,
-      "accuracy": 0.6924,
-      "speaker_agreement": 0.8333
+      "jensen_shannon": 0.011022,
+      "cosine_similarity": 0.999562,
+      "kl_divergence": 0.000486,
+      "total_variation": 0.014378,
+      "n": 759,
+      "accuracy": 0.7141,
+      "speaker_agreement": 0.875
     }
   ],
   "gt_reference_distribution": {
-    "Mild": 0.0,
-    "Moderate": 0.5,
+    "Mild": 0.0875,
+    "Moderate": 0.25,
     "Normal": 0.325,
-    "Severe": 0.175
+    "Severe": 0.3375
   },
   "severity_labels": [
     "Mild",