glecorve commited on
Commit
ffd1156
β€’
1 Parent(s): f31a098

Enriched README

Browse files
Files changed (1) hide show
  1. README.md +172 -3
README.md CHANGED
@@ -1,6 +1,4 @@
1
  ---
2
- language:
3
- - en
4
  dataset_info:
5
  features:
6
  - name: original_nl_question
@@ -41,6 +39,177 @@ dataset_info:
41
  num_examples: 9961
42
  download_size: 7595264
43
  dataset_size: 16322261
 
 
 
 
 
 
 
 
 
44
  ---
45
 
46
- # Dataset Card for SimpleQuestions-SPARQLtoText
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
 
 
2
  dataset_info:
3
  features:
4
  - name: original_nl_question
 
39
  num_examples: 9961
40
  download_size: 7595264
41
  dataset_size: 16322261
42
+ task_categories:
43
+ - question-answering
44
+ - text-generation
45
+ tags:
46
+ - qa
47
+ - knowledge-graph
48
+ - sparql
49
+ language:
50
+ - en
51
  ---
52
 
53
+ # Dataset Card for SimpleQuestions-SPARQLtoText
54
+
55
+ ## Table of Contents
56
+ - [Dataset Card for SimpleQuestions-SPARQLtoText](#dataset-card-for-simplequestions-sparqltotext)
57
+ - [Table of Contents](#table-of-contents)
58
+ - [Dataset Description](#dataset-description)
59
+ - [Dataset Summary](#dataset-summary)
60
+ - [JSON fields](#json-fields)
61
+ - [Format of the SPARQL queries](#format-of-the-sparql-queries)
62
+ - [Answerable/unanswerable](#answerableunanswerable)
63
+ - [Languages](#languages)
64
+ - [Dataset Structure](#dataset-structure)
65
+ - [Types of questions](#types-of-questions)
66
+ - [Data splits](#data-splits)
67
+ - [Additional information](#additional-information)
68
+ - [Related datasets](#related-datasets)
69
+ - [Licencing information](#licencing-information)
70
+ - [Citation information](#citation-information)
71
+ - [This version of the corpus (with normalized SPARQL queries)](#this-version-of-the-corpus-with-normalized-sparql-queries)
72
+ - [Original version](#original-version)
73
+
74
+
75
+ ## Dataset Description
76
+
77
+ - **Paper:** [SPARQL-to-Text Question Generation for Knowledge-Based Conversational Applications (AACL-IJCNLP 2022)](https://aclanthology.org/2022.aacl-main.11/)
78
+ - **Point of Contact:** GwΓ©nolΓ© LecorvΓ©
79
+
80
+ ### Dataset Summary
81
+
82
+ Special version of [SimpleQuestions](https://github.com/askplatypus/wikidata-simplequestions) with SPARQL queries formatted for the SPARQL-to-Text task.
83
+
84
+ #### JSON fields
85
+
86
+ The original version of SimpleQuestions is a raw text file listing triples and the natural language question. A JSON version has been generated and augmented with the following fields:
87
+
88
+ * `rdf_subject`, `rdf_property`, `rdf_object`: triple in the Wikidata format (IDs)
89
+
90
+ * `nl_subject`, `nl_property`, `nl_object`: triple with labels retrieved from Wikidata. Some entities do not have labels, they are labelled as `UNDEFINED_LABEL`
91
+
92
+ * `sparql_query`: SPARQL query with Wikidata IDs
93
+
94
+ * `verbalized_sparql_query`: SPARQL query with labels
95
+
96
+ * `original_nl_question`: original natural language question from SimpleQuestions. This is in **lower case**.
97
+
98
+ * `recased_nl_question`: Version of `original_nl_question` where the named entities have been automatically recased based on the labels of the entities.
99
+
100
+ #### Format of the SPARQL queries
101
+
102
+ * Randomizing the variables names
103
+
104
+ * Delimiters are spaced
105
+
106
+ #### Answerable/unanswerable
107
+
108
+ Some questions in SimpleQuestions cannot be answered. Hence, it originally comes with 2 versions for the train/valid/test sets: one with all entries, another with the answerable questions only.
109
+
110
+ ### Languages
111
+
112
+ - English
113
+
114
+ ## Dataset Structure
115
+
116
+ ### Types of questions
117
+
118
+ Comparison of question types compared to related datasets:
119
+
120
+ | | | [SimpleQuestions](https://huggingface.co/datasets/OrangeInnov/simplequestions-sparqltotext) | [ParaQA](https://huggingface.co/datasets/OrangeInnov/paraqa-sparqltotext) | [LC-QuAD 2.0](https://huggingface.co/datasets/OrangeInnov/lcquad_2.0-sparqltotext) | [CSQA](https://huggingface.co/datasets/OrangeInnov/csqa-sparqltotext) | [WebNLQ-QA](https://huggingface.co/datasets/OrangeInnov/webnlg-qa) |
121
+ |--------------------------|-----------------|:---------------:|:------:|:-----------:|:----:|:---------:|
122
+ | **Number of triplets in query** | 1 | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
123
+ | | 2 | | βœ“ | βœ“ | βœ“ | βœ“ |
124
+ | | More | | | βœ“ | βœ“ | βœ“ |
125
+ | **Logical connector between triplets** | Conjunction | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
126
+ | | Disjunction | | | | βœ“ | βœ“ |
127
+ | | Exclusion | | | | βœ“ | βœ“ |
128
+ | **Topology of the query graph** | Direct | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
129
+ | | Sibling | | βœ“ | βœ“ | βœ“ | βœ“ |
130
+ | | Chain | | βœ“ | βœ“ | βœ“ | βœ“ |
131
+ | | Mixed | | | βœ“ | | βœ“ |
132
+ | | Other | | βœ“ | βœ“ | βœ“ | βœ“ |
133
+ | **Variable typing in the query** | None | βœ“ | βœ“ | βœ“ | βœ“ | οΏ½οΏ½ |
134
+ | | Target variable | | βœ“ | βœ“ | βœ“ | βœ“ |
135
+ | | Internal variable | | βœ“ | βœ“ | βœ“ | βœ“ |
136
+ | **Comparisons clauses** | None | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
137
+ | | String | | | βœ“ | | βœ“ |
138
+ | | Number | | | βœ“ | βœ“ | βœ“ |
139
+ | | Date | | | βœ“ | | βœ“ |
140
+ | **Superlative clauses** | No | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
141
+ | | Yes | | | | βœ“ | |
142
+ | **Answer type** | Entity (open) | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
143
+ | | Entity (closed) | | | | βœ“ | βœ“ |
144
+ | | Number | | | βœ“ | βœ“ | βœ“ |
145
+ | | Boolean | | βœ“ | βœ“ | βœ“ | βœ“ |
146
+ | **Answer cardinality** | 0 (unanswerable) | | | βœ“ | | βœ“ |
147
+ | | 1 | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
148
+ | | More | | βœ“ | βœ“ | βœ“ | βœ“ |
149
+ | **Number of target variables** | 0 (β‡’ ASK verb) | | βœ“ | βœ“ | βœ“ | βœ“ |
150
+ | | 1 | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
151
+ | | 2 | | | βœ“ | | βœ“ |
152
+ | **Dialogue context** | Self-sufficient | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
153
+ | | Coreference | | | | βœ“ | βœ“ |
154
+ | | Ellipsis | | | | βœ“ | βœ“ |
155
+ | **Meaning** | Meaningful | βœ“ | βœ“ | βœ“ | βœ“ | βœ“ |
156
+ | | Non-sense | | | | | βœ“ |
157
+
158
+
159
+ ### Data splits
160
+
161
+ Text verbalization is only available for a subset of the test set, referred to as *challenge set*. Other sample only contain dialogues in the form of follow-up sparql queries.
162
+
163
+ | | Train | Validation | Test |
164
+ | --------------------- | ---------- | ---------- | ---------- |
165
+ | Questions | 34,000 | 5,000 | 10,000 |
166
+ | NL question per query | 1 |
167
+ | Characters per query | 70 (Β± 10) |
168
+ | Tokens per question | 7.4 (Β± 2.1) |
169
+
170
+
171
+ ## Additional information
172
+
173
+ ### Related datasets
174
+
175
+ This corpus is part of a set of 5 datasets released for SPARQL-to-Text generation, namely:
176
+ - Non conversational datasets
177
+ - [SimpleQuestions](https://huggingface.co/datasets/OrangeInnov/simplequestions-sparqltotext) (from https://github.com/askplatypus/wikidata-simplequestions)
178
+ - [ParaQA](https://huggingface.co/datasets/OrangeInnov/paraqa-sparqltotext) (from https://github.com/barshana-banerjee/ParaQA)
179
+ - [LC-QuAD 2.0](https://huggingface.co/datasets/OrangeInnov/lcquad_2.0-sparqltotext) (from http://lc-quad.sda.tech/)
180
+ - Conversational datasets
181
+ - [CSQA](https://huggingface.co/datasets/OrangeInnov/csqa-sparqltotext) (from https://amritasaha1812.github.io/CSQA/)
182
+ - [WebNLQ-QA](https://huggingface.co/datasets/OrangeInnov/webnlg-qa) (derived from https://gitlab.com/shimorina/webnlg-dataset/-/tree/master/release_v3.0)
183
+
184
+ ### Licencing information
185
+
186
+ * Content from original dataset: CC-BY 3.0
187
+ * New content: CC BY-SA 4.0
188
+
189
+
190
+
191
+ ### Citation information
192
+
193
+
194
+ #### This version of the corpus (with normalized SPARQL queries)
195
+
196
+ ```bibtex
197
+ @inproceedings{lecorve2022sparql2text,
198
+ title={SPARQL-to-Text Question Generation for Knowledge-Based Conversational Applications},
199
+ author={Lecorv\'e, Gw\'enol\'e and Veyret, Morgan and Brabant, Quentin and Rojas-Barahona, Lina M.},
200
+ journal={Proceedings of the Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the International Joint Conference on Natural Language Processing (AACL-IJCNLP)},
201
+ year={2022}
202
+ }
203
+ ```
204
+
205
+ #### Original version
206
+
207
+ ```bibtex
208
+ @article{bordes2015large,
209
+ title={Large-scale simple question answering with memory networks},
210
+ author={Bordes, Antoine and Usunier, Nicolas and Chopra, Sumit and Weston, Jason},
211
+ journal={arXiv preprint arXiv:1506.02075},
212
+ year={2015}
213
+ }
214
+
215
+ ```