Datasets:
Tasks:
Text Classification
Modalities:
Text
Sub-tasks:
natural-language-inference
Size:
100K - 1M
Tags:
quality-estimation
License:
Finished README
Browse files
README.md
CHANGED
@@ -32,6 +32,7 @@ task_ids:
|
|
32 |
- [Dataset Structure](#dataset-structure)
|
33 |
- [Data Instances](#data-instances)
|
34 |
- [Data Splits](#data-splits)
|
|
|
35 |
- [Dataset Creation](#dataset-creation)
|
36 |
- [Additional Information](#additional-information)
|
37 |
- [Dataset Curators](#dataset-curators)
|
@@ -51,49 +52,34 @@ The intended usage of this corpus is restricted to the scope of final project fo
|
|
51 |
The language data of this corpus is in English (BCP-47 `en`) and Dutch (BCP-47 `nl`).
|
52 |
## Dataset Structure
|
53 |
### Data Instances
|
54 |
-
The dataset contains a single configuration, named `plain_text`, containing the following fields:
|
55 |
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
81 |
-
|
82 |
-
|
83 |
-
|
84 |
-
- `da_explanation_2`: The quality estimation produced by the `wmt20-comet-qe-da` model for the second explanation translation.
|
85 |
-
|
86 |
-
- `da_explanation_3`: The quality estimation produced by the `wmt20-comet-qe-da` model for the third explanation translation.
|
87 |
-
|
88 |
-
- `mqm_premise`: The quality estimation produced by the `wmt21-comet-qe-mqm` model for the premise translation.
|
89 |
-
|
90 |
-
- `mqm_hypothesis`: The quality estimation produced by the `wmt21-comet-qe-mqm` model for the hypothesis translation.
|
91 |
-
|
92 |
-
- `mqm_explanation_1`: The quality estimation produced by the `wmt21-comet-qe-mqm` model for the first explanation translation.
|
93 |
-
|
94 |
-
- `mqm_explanation_2`: The quality estimation produced by the `wmt21-comet-qe-mqm` model for the second explanation translation.
|
95 |
-
|
96 |
-
- `mqm_explanation_3`: The quality estimation produced by the `wmt21-comet-qe-mqm` model for the third explanation translation.
|
97 |
|
98 |
### Data Splits
|
99 |
|
@@ -103,6 +89,36 @@ The dataset contains a single configuration, named `plain_text`, containing the
|
|
103 |
|
104 |
For your analyses, use the amount of data that is the most reasonable for your computational setup. The more, the better.
|
105 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
106 |
### Dataset Creation
|
107 |
|
108 |
The dataset was created through the following steps:
|
|
|
32 |
- [Dataset Structure](#dataset-structure)
|
33 |
- [Data Instances](#data-instances)
|
34 |
- [Data Splits](#data-splits)
|
35 |
+
- [Data Example](#data-example)
|
36 |
- [Dataset Creation](#dataset-creation)
|
37 |
- [Additional Information](#additional-information)
|
38 |
- [Dataset Curators](#dataset-curators)
|
|
|
52 |
The language data of this corpus is in English (BCP-47 `en`) and Dutch (BCP-47 `nl`).
|
53 |
## Dataset Structure
|
54 |
### Data Instances
|
|
|
55 |
|
56 |
+
The dataset contains a single condiguration by default, named `plain_text`, with the three original splits `train`, `validation` and `test`. Every split contains the following fields:
|
57 |
+
|
58 |
+
| **Field** | **Description** |
|
59 |
+
|------------|-----------------------------|
|
60 |
+
|`premise_en`| The original English premise.|
|
61 |
+
|`premise_nl`| The premise automatically translated to Dutch.|
|
62 |
+
|`hypothesis_en`| The original English hypothesis.|
|
63 |
+
|`hypothesis_nl`| The hypothesis automatically translated to Dutch.|
|
64 |
+
|`label`| The label of the data instance (0 for entailment, 1 for neutral, 2 for contradiction).|
|
65 |
+
|`explanation_1_en`| The first explanation for the assigned label in English.|
|
66 |
+
|`explanation_1_nl`| The first explanation automatically translated to Dutch.|
|
67 |
+
|`explanation_2_en`| The second explanation for the assigned label in English.|
|
68 |
+
|`explanation_2_nl`| The second explanation automatically translated to Dutch.|
|
69 |
+
|`explanation_3_en`| The third explanation for the assigned label in English.|
|
70 |
+
|`explanation_3_nl`| The third explanation automatically translated to Dutch.|
|
71 |
+
|`da_premise`| The quality estimation produced by the `wmt20-comet-qe-da` model for the premise translation.|
|
72 |
+
|`da_hypothesis`| The quality estimation produced by the `wmt20-comet-qe-da` model for the hypothesis translation.|
|
73 |
+
|`da_explanation_1`| The quality estimation produced by the `wmt20-comet-qe-da` model for the first explanation translation.|
|
74 |
+
|`da_explanation_2`| The quality estimation produced by the `wmt20-comet-qe-da` model for the second explanation translation.|
|
75 |
+
|`da_explanation_3`| The quality estimation produced by the `wmt20-comet-qe-da` model for the third explanation translation.|
|
76 |
+
|`mqm_premise`| The quality estimation produced by the `wmt21-comet-qe-mqm` model for the premise translation.|
|
77 |
+
|`mqm_hypothesis`| The quality estimation produced by the `wmt21-comet-qe-mqm` model for the hypothesis translation.|
|
78 |
+
|`mqm_explanation_1`| The quality estimation produced by the `wmt21-comet-qe-mqm` model for the first explanation translation.|
|
79 |
+
|`mqm_explanation_2`| The quality estimation produced by the `wmt21-comet-qe-mqm` model for the second explanation translation.|
|
80 |
+
|`mqm_explanation_3`| The quality estimation produced by the `wmt21-comet-qe-mqm` model for the third explanation translation.|
|
81 |
+
|
82 |
+
Explanation 2 and 3 and related quality estimation scores are only present in the `validation` and `test` splits.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
83 |
|
84 |
### Data Splits
|
85 |
|
|
|
89 |
|
90 |
For your analyses, use the amount of data that is the most reasonable for your computational setup. The more, the better.
|
91 |
|
92 |
+
### Data Example
|
93 |
+
|
94 |
+
The following is an example of entry 2000 taken from the `test` split:
|
95 |
+
|
96 |
+
```json
|
97 |
+
{
|
98 |
+
"premise_en": "A young woman wearing a yellow sweater and black pants is ice skating outdoors.",
|
99 |
+
"premise_nl": "Een jonge vrouw met een gele trui en zwarte broek schaatst buiten.",
|
100 |
+
"hypothesis_en": "a woman is practicing for the olympics",
|
101 |
+
"hypothesis_nl": "een vrouw oefent voor de Olympische Spelen",
|
102 |
+
"label": 1,
|
103 |
+
"explanation_1_en": "You can not infer it's for the Olympics.",
|
104 |
+
"explanation_1_nl": "Het is niet voor de Olympische Spelen.",
|
105 |
+
"explanation_2_en": "Just because a girl is skating outdoors does not mean she is practicing for the Olympics.",
|
106 |
+
"explanation_2_nl": "Alleen omdat een meisje buiten schaatst betekent niet dat ze oefent voor de Olympische Spelen.",
|
107 |
+
"explanation_3_en": "Ice skating doesn't imply practicing for the olympics.",
|
108 |
+
"explanation_3_nl": "Schaatsen betekent niet oefenen voor de Olympische Spelen.",
|
109 |
+
"da_premise": "0.6099",
|
110 |
+
"mqm_premise": "0.1298",
|
111 |
+
"da_hypothesis": "0.8504",
|
112 |
+
"mqm_hypothesis": "0.1521",
|
113 |
+
"da_explanation_1": "0.0001",
|
114 |
+
"mqm_explanation_1": "0.1237",
|
115 |
+
"da_explanation_2": "0.4017",
|
116 |
+
"mqm_explanation_2": "0.1467",
|
117 |
+
"da_explanation_3": "0.6069",
|
118 |
+
"mqm_explanation_3": "0.1389"
|
119 |
+
}
|
120 |
+
```
|
121 |
+
|
122 |
### Dataset Creation
|
123 |
|
124 |
The dataset was created through the following steps:
|