Convert dataset to Parquet

Convert dataset to Parquet.

README.md

@@ -20,22 +20,6 @@ task_ids:
 - open-domain-qa
 paperswithcode_id: adversarialqa
 pretty_name: adversarialQA
-train-eval-index:
-- config: adversarialQA
-  task: question-answering
-  task_id: extractive_question_answering
-  splits:
-    train_split: train
-    eval_split: validation
-  col_mapping:
-    question: question
-    context: context
-    answers:
-      text: text
-      answer_start: answer_start
-  metrics:
-  - type: squad
-    name: SQuAD
 dataset_info:
 - config_name: adversarialQA
   features:
@@ -61,17 +45,17 @@ dataset_info:
       dtype: string
   splits:
   - name: train
-    num_bytes: 27858794
+    num_bytes: 27858686
     num_examples: 30000
   - name: validation
-    num_bytes: 2757128
+    num_bytes: 2757092
     num_examples: 3000
   - name: test
-    num_bytes: 2919643
+    num_bytes: 2919479
     num_examples: 3000
-  download_size: 9018914
-  dataset_size: 33535565
-- config_name: dbidaf
+  download_size: 5301049
+  dataset_size: 33535257
+- config_name: dbert
   features:
   - name: id
     dtype: string
@@ -95,17 +79,17 @@ dataset_info:
       dtype: string
   splits:
   - name: train
-    num_bytes: 9282518
+    num_bytes: 9345557
     num_examples: 10000
   - name: validation
-    num_bytes: 917943
+    num_bytes: 918192
     num_examples: 1000
   - name: test
-    num_bytes: 947111
+    num_bytes: 971454
     num_examples: 1000
   download_size: 9018914
-  dataset_size: 11147572
-- config_name: dbert
+  dataset_size: 11235203
+- config_name: dbidaf
   features:
   - name: id
     dtype: string
@@ -129,16 +113,16 @@ dataset_info:
       dtype: string
   splits:
   - name: train
-    num_bytes: 9345557
+    num_bytes: 9282518
     num_examples: 10000
   - name: validation
-    num_bytes: 918192
+    num_bytes: 917943
     num_examples: 1000
   - name: test
-    num_bytes: 971454
+    num_bytes: 947111
     num_examples: 1000
   download_size: 9018914
-  dataset_size: 11235203
+  dataset_size: 11147572
 - config_name: droberta
   features:
   - name: id
@@ -173,6 +157,31 @@ dataset_info:
     num_examples: 1000
   download_size: 9018914
   dataset_size: 11201190
+configs:
+- config_name: adversarialQA
+  data_files:
+  - split: train
+    path: adversarialQA/train-*
+  - split: validation
+    path: adversarialQA/validation-*
+  - split: test
+    path: adversarialQA/test-*
+train-eval-index:
+- config: adversarialQA
+  task: question-answering
+  task_id: extractive_question_answering
+  splits:
+    train_split: train
+    eval_split: validation
+  col_mapping:
+    question: question
+    context: context
+    answers:
+      text: text
+      answer_start: answer_start
+  metrics:
+  - type: squad
+    name: SQuAD
 ---
 
 # Dataset Card for adversarialQA

adversarialQA/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a84cab6ed591ed6024c9eab388131fb40cfcd01b0a76d84735b85115b4bde76
+size 457062

adversarialQA/train-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0010a888ffe67a4da4afb7a9cbb7e0e77daaf02aaf055ee1697c88519afb94d8
+size 4348721

adversarialQA/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ed5569eb1dedc73ad7a1dc34d9064ca8693216b1b47e5c2773a9b2a81668f1f
+size 495266

dataset_infos.json

@@ -1,394 +1,375 @@
 {
-  "adversarialQA": {
-    "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
-    "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
-    "homepage": "https://adversarialqa.github.io/",
-    "license": "",
-    "features": {
-      "id": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "title": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "context": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "question": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "answers": {
-        "feature": {
-          "text": {
-            "dtype": "string",
-            "id": null,
-            "_type": "Value"
-          },
-          "answer_start": {
-            "dtype": "int32",
-            "id": null,
-            "_type": "Value"
-          }
+    "adversarialQA": {
+        "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
+        "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl_a_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
+        "homepage": "https://adversarialqa.github.io/",
+        "license": "",
+        "features": {
+            "id": {
+                "dtype": "string",
+                "_type": "Value"
+            },
+            "title": {
+                "dtype": "string",
+                "_type": "Value"
+            },
+            "context": {
+                "dtype": "string",
+                "_type": "Value"
+            },
+            "question": {
+                "dtype": "string",
+                "_type": "Value"
+            },
+            "answers": {
+                "feature": {
+                    "text": {
+                        "dtype": "string",
+                        "_type": "Value"
+                    },
+                    "answer_start": {
+                        "dtype": "int32",
+                        "_type": "Value"
+                    }
+                },
+                "_type": "Sequence"
+            },
+            "metadata": {
+                "split": {
+                    "dtype": "string",
+                    "_type": "Value"
+                },
+                "model_in_the_loop": {
+                    "dtype": "string",
+                    "_type": "Value"
+                }
+            }
         },
-        "length": -1,
-        "id": null,
-        "_type": "Sequence"
-      },
-      "metadata": {
-        "split": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
+        "builder_name": "adversarial_qa",
+        "dataset_name": "adversarial_qa",
+        "config_name": "adversarialQA",
+        "version": {
+            "version_str": "1.0.0",
+            "major": 1,
+            "minor": 0,
+            "patch": 0
         },
-        "model_in_the_loop": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
-        }
-      }
-    },
-    "post_processed": null,
-    "supervised_keys": null,
-    "builder_name": "adversarial_qa",
-    "config_name": "adversarialQA",
-    "version": {
-      "version_str": "1.0.0",
-      "description": null,
-      "major": 1,
-      "minor": 0,
-      "patch": 0
-    },
-    "splits": {
-      "train": {
-        "name": "train",
-        "num_bytes": 27858794,
-        "num_examples": 30000,
-        "dataset_name": "adversarial_qa"
-      },
-      "validation": {
-        "name": "validation",
-        "num_bytes": 2757128,
-        "num_examples": 3000,
-        "dataset_name": "adversarial_qa"
-      },
-      "test": {
-        "name": "test",
-        "num_bytes": 2919643,
-        "num_examples": 3000,
-        "dataset_name": "adversarial_qa"
-      }
-    },
-    "download_checksums": {
-      "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
-        "num_bytes": 9018914,
-        "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
-      }
+        "splits": {
+            "train": {
+                "name": "train",
+                "num_bytes": 27858686,
+                "num_examples": 30000,
+                "dataset_name": null
+            },
+            "validation": {
+                "name": "validation",
+                "num_bytes": 2757092,
+                "num_examples": 3000,
+                "dataset_name": null
+            },
+            "test": {
+                "name": "test",
+                "num_bytes": 2919479,
+                "num_examples": 3000,
+                "dataset_name": null
+            }
+        },
+        "download_size": 5301049,
+        "dataset_size": 33535257,
+        "size_in_bytes": 38836306
     },
-    "download_size": 9018914,
-    "post_processing_size": null,
-    "dataset_size": 33535565,
-    "size_in_bytes": 42554479
-  },
-  "dbidaf": {
-    "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
-    "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
-    "homepage": "https://adversarialqa.github.io/",
-    "license": "",
-    "features": {
-      "id": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "title": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "context": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "question": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "answers": {
-        "feature": {
-          "text": {
-            "dtype": "string",
-            "id": null,
-            "_type": "Value"
-          },
-          "answer_start": {
-            "dtype": "int32",
-            "id": null,
-            "_type": "Value"
-          }
+    "dbidaf": {
+        "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
+        "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
+        "homepage": "https://adversarialqa.github.io/",
+        "license": "",
+        "features": {
+            "id": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "title": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "context": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "question": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "answers": {
+                "feature": {
+                    "text": {
+                        "dtype": "string",
+                        "id": null,
+                        "_type": "Value"
+                    },
+                    "answer_start": {
+                        "dtype": "int32",
+                        "id": null,
+                        "_type": "Value"
+                    }
+                },
+                "length": -1,
+                "id": null,
+                "_type": "Sequence"
+            },
+            "metadata": {
+                "split": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                },
+                "model_in_the_loop": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                }
+            }
         },
-        "length": -1,
-        "id": null,
-        "_type": "Sequence"
-      },
-      "metadata": {
-        "split": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
+        "post_processed": null,
+        "supervised_keys": null,
+        "builder_name": "adversarial_qa",
+        "config_name": "dbidaf",
+        "version": {
+            "version_str": "1.0.0",
+            "description": null,
+            "major": 1,
+            "minor": 0,
+            "patch": 0
         },
-        "model_in_the_loop": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
-        }
-      }
-    },
-    "post_processed": null,
-    "supervised_keys": null,
-    "builder_name": "adversarial_qa",
-    "config_name": "dbidaf",
-    "version": {
-      "version_str": "1.0.0",
-      "description": null,
-      "major": 1,
-      "minor": 0,
-      "patch": 0
-    },
-    "splits": {
-      "train": {
-        "name": "train",
-        "num_bytes": 9282518,
-        "num_examples": 10000,
-        "dataset_name": "adversarial_qa"
-      },
-      "validation": {
-        "name": "validation",
-        "num_bytes": 917943,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      },
-      "test": {
-        "name": "test",
-        "num_bytes": 947111,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      }
-    },
-    "download_checksums": {
-      "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
-        "num_bytes": 9018914,
-        "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
-      }
-    },
-    "download_size": 9018914,
-    "post_processing_size": null,
-    "dataset_size": 11147572,
-    "size_in_bytes": 20166486
-  },
-  "dbert": {
-    "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
-    "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
-    "homepage": "https://adversarialqa.github.io/",
-    "license": "",
-    "features": {
-      "id": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "title": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "context": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "question": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "answers": {
-        "feature": {
-          "text": {
-            "dtype": "string",
-            "id": null,
-            "_type": "Value"
-          },
-          "answer_start": {
-            "dtype": "int32",
-            "id": null,
-            "_type": "Value"
-          }
+        "splits": {
+            "train": {
+                "name": "train",
+                "num_bytes": 9282518,
+                "num_examples": 10000,
+                "dataset_name": "adversarial_qa"
+            },
+            "validation": {
+                "name": "validation",
+                "num_bytes": 917943,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            },
+            "test": {
+                "name": "test",
+                "num_bytes": 947111,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            }
         },
-        "length": -1,
-        "id": null,
-        "_type": "Sequence"
-      },
-      "metadata": {
-        "split": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
+        "download_checksums": {
+            "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
+                "num_bytes": 9018914,
+                "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
+            }
         },
-        "model_in_the_loop": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
-        }
-      }
-    },
-    "post_processed": null,
-    "supervised_keys": null,
-    "builder_name": "adversarial_qa",
-    "config_name": "dbert",
-    "version": {
-      "version_str": "1.0.0",
-      "description": null,
-      "major": 1,
-      "minor": 0,
-      "patch": 0
-    },
-    "splits": {
-      "train": {
-        "name": "train",
-        "num_bytes": 9345557,
-        "num_examples": 10000,
-        "dataset_name": "adversarial_qa"
-      },
-      "validation": {
-        "name": "validation",
-        "num_bytes": 918192,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      },
-      "test": {
-        "name": "test",
-        "num_bytes": 971454,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      }
+        "download_size": 9018914,
+        "post_processing_size": null,
+        "dataset_size": 11147572,
+        "size_in_bytes": 20166486
     },
-    "download_checksums": {
-      "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
-        "num_bytes": 9018914,
-        "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
-      }
-    },
-    "download_size": 9018914,
-    "post_processing_size": null,
-    "dataset_size": 11235203,
-    "size_in_bytes": 20254117
-  },
-  "droberta": {
-    "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
-    "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
-    "homepage": "https://adversarialqa.github.io/",
-    "license": "",
-    "features": {
-      "id": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "title": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "context": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "question": {
-        "dtype": "string",
-        "id": null,
-        "_type": "Value"
-      },
-      "answers": {
-        "feature": {
-          "text": {
-            "dtype": "string",
-            "id": null,
-            "_type": "Value"
-          },
-          "answer_start": {
-            "dtype": "int32",
-            "id": null,
-            "_type": "Value"
-          }
+    "dbert": {
+        "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
+        "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
+        "homepage": "https://adversarialqa.github.io/",
+        "license": "",
+        "features": {
+            "id": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "title": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "context": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "question": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "answers": {
+                "feature": {
+                    "text": {
+                        "dtype": "string",
+                        "id": null,
+                        "_type": "Value"
+                    },
+                    "answer_start": {
+                        "dtype": "int32",
+                        "id": null,
+                        "_type": "Value"
+                    }
+                },
+                "length": -1,
+                "id": null,
+                "_type": "Sequence"
+            },
+            "metadata": {
+                "split": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                },
+                "model_in_the_loop": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                }
+            }
         },
-        "length": -1,
-        "id": null,
-        "_type": "Sequence"
-      },
-      "metadata": {
-        "split": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
+        "post_processed": null,
+        "supervised_keys": null,
+        "builder_name": "adversarial_qa",
+        "config_name": "dbert",
+        "version": {
+            "version_str": "1.0.0",
+            "description": null,
+            "major": 1,
+            "minor": 0,
+            "patch": 0
         },
-        "model_in_the_loop": {
-          "dtype": "string",
-          "id": null,
-          "_type": "Value"
-        }
-      }
-    },
-    "post_processed": null,
-    "supervised_keys": null,
-    "builder_name": "adversarial_qa",
-    "config_name": "droberta",
-    "version": {
-      "version_str": "1.0.0",
-      "description": null,
-      "major": 1,
-      "minor": 0,
-      "patch": 0
-    },
-    "splits": {
-      "train": {
-        "name": "train",
-        "num_bytes": 9270719,
-        "num_examples": 10000,
-        "dataset_name": "adversarial_qa"
-      },
-      "validation": {
-        "name": "validation",
-        "num_bytes": 925065,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      },
-      "test": {
-        "name": "test",
-        "num_bytes": 1005406,
-        "num_examples": 1000,
-        "dataset_name": "adversarial_qa"
-      }
-    },
-    "download_checksums": {
-      "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
-        "num_bytes": 9018914,
-        "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
-      }
+        "splits": {
+            "train": {
+                "name": "train",
+                "num_bytes": 9345557,
+                "num_examples": 10000,
+                "dataset_name": "adversarial_qa"
+            },
+            "validation": {
+                "name": "validation",
+                "num_bytes": 918192,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            },
+            "test": {
+                "name": "test",
+                "num_bytes": 971454,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            }
+        },
+        "download_checksums": {
+            "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
+                "num_bytes": 9018914,
+                "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
+            }
+        },
+        "download_size": 9018914,
+        "post_processing_size": null,
+        "dataset_size": 11235203,
+        "size_in_bytes": 20254117
     },
-    "download_size": 9018914,
-    "post_processing_size": null,
-    "dataset_size": 11201190,
-    "size_in_bytes": 20220104
-  }
+    "droberta": {
+        "description": "AdversarialQA is a Reading Comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles using an adversarial model-in-the-loop.\nWe use three different models; BiDAF (Seo et al., 2016), BERT-Large (Devlin et al., 2018), and RoBERTa-Large (Liu et al., 2019) in the annotation loop and construct three datasets; D(BiDAF), D(BERT), and D(RoBERTa), each with 10,000 training examples, 1,000 validation, and 1,000 test examples.\nThe adversarial human annotation paradigm ensures that these datasets consist of questions that current state-of-the-art models (at least the ones used as adversaries in the annotation loop) find challenging.\n",
+        "citation": "@article{bartolo2020beat,\n    author = {Bartolo, Max and Roberts, Alastair and Welbl, Johannes and Riedel, Sebastian and Stenetorp, Pontus},\n    title = {Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehension},\n    journal = {Transactions of the Association for Computational Linguistics},\n    volume = {8},\n    number = {},\n    pages = {662-678},\n    year = {2020},\n    doi = {10.1162/tacl\\_a\\_00338},\n    URL = { https://doi.org/10.1162/tacl_a_00338 },\n    eprint = { https://doi.org/10.1162/tacl_a_00338 },\n    abstract = { Innovations in annotation methodology have been a catalyst for Reading Comprehension (RC) datasets and models. One recent trend to challenge current RC models is to involve a model in the annotation process: Humans create questions adversarially, such that the model fails to answer them correctly. In this work we investigate this annotation methodology and apply it in three different settings, collecting a total of 36,000 samples with progressively stronger models in the annotation loop. This allows us to explore questions such as the reproducibility of the adversarial effect, transfer from data collected with varying model-in-the-loop strengths, and generalization to data collected without a model. We find that training on adversarially collected samples leads to strong generalization to non-adversarially collected datasets, yet with progressive performance deterioration with increasingly stronger models-in-the-loop. Furthermore, we find that stronger models can still learn from datasets collected with substantially weaker models-in-the-loop. When trained on data collected with a BiDAF model in the loop, RoBERTa achieves 39.9F1 on questions that it cannot answer when trained on SQuAD\u2014only marginally lower than when trained on data collected using RoBERTa itself (41.0F1). }\n}\n",
+        "homepage": "https://adversarialqa.github.io/",
+        "license": "",
+        "features": {
+            "id": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "title": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "context": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "question": {
+                "dtype": "string",
+                "id": null,
+                "_type": "Value"
+            },
+            "answers": {
+                "feature": {
+                    "text": {
+                        "dtype": "string",
+                        "id": null,
+                        "_type": "Value"
+                    },
+                    "answer_start": {
+                        "dtype": "int32",
+                        "id": null,
+                        "_type": "Value"
+                    }
+                },
+                "length": -1,
+                "id": null,
+                "_type": "Sequence"
+            },
+            "metadata": {
+                "split": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                },
+                "model_in_the_loop": {
+                    "dtype": "string",
+                    "id": null,
+                    "_type": "Value"
+                }
+            }
+        },
+        "post_processed": null,
+        "supervised_keys": null,
+        "builder_name": "adversarial_qa",
+        "config_name": "droberta",
+        "version": {
+            "version_str": "1.0.0",
+            "description": null,
+            "major": 1,
+            "minor": 0,
+            "patch": 0
+        },
+        "splits": {
+            "train": {
+                "name": "train",
+                "num_bytes": 9270719,
+                "num_examples": 10000,
+                "dataset_name": "adversarial_qa"
+            },
+            "validation": {
+                "name": "validation",
+                "num_bytes": 925065,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            },
+            "test": {
+                "name": "test",
+                "num_bytes": 1005406,
+                "num_examples": 1000,
+                "dataset_name": "adversarial_qa"
+            }
+        },
+        "download_checksums": {
+            "https://adversarialqa.github.io/data/aqa_v1.0.zip": {
+                "num_bytes": 9018914,
+                "checksum": "f4f3c23224a5060b28c35e35581bd5cf46256dda3665418fb83d036d0e0c93cf"
+            }
+        },
+        "download_size": 9018914,
+        "post_processing_size": null,
+        "dataset_size": 11201190,
+        "size_in_bytes": 20220104
+    }
 }