mveyret commited on
Commit
e4fb397
0 Parent(s):

initial commit

Browse files
.gitattributes ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ dstc2-rdf/test.jsonl filter=lfs diff=lfs merge=lfs -text
2
+ multiwoz-rdf/train.jsonl filter=lfs diff=lfs merge=lfs -text
3
+ multiwoz-sim-rdf/validation.jsonl filter=lfs diff=lfs merge=lfs -text
4
+ camrest-sim-rdf/test.jsonl filter=lfs diff=lfs merge=lfs -text
5
+ camrest-sim-rdf/train.jsonl filter=lfs diff=lfs merge=lfs -text
6
+ camrest-sim-rdf/validation.jsonl filter=lfs diff=lfs merge=lfs -text
7
+ sfxdial-rdf/train.jsonl filter=lfs diff=lfs merge=lfs -text
8
+ sfxdial-rdf/validation.jsonl filter=lfs diff=lfs merge=lfs -text
9
+ dstc2-rdf/train.jsonl filter=lfs diff=lfs merge=lfs -text
10
+ dstc2-rdf/validation.jsonl filter=lfs diff=lfs merge=lfs -text
11
+ multiwoz-rdf/validation.jsonl filter=lfs diff=lfs merge=lfs -text
12
+ multiwoz-sim-rdf/train.jsonl filter=lfs diff=lfs merge=lfs -text
13
+ multiwoz-rdf/test.jsonl filter=lfs diff=lfs merge=lfs -text
14
+ multiwoz-sim-rdf/test.jsonl filter=lfs diff=lfs merge=lfs -text
15
+ sfxdial-rdf/test.jsonl filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,291 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ configs:
3
+ - config_name: bundle-converted
4
+ description: Merge of all rdf converted datasets
5
+ data_files:
6
+ - path: ["dstc2-rdf/train.jsonl","multiwoz-rdf/train.jsonl","sfxdial-rdf/train.jsonl"]
7
+ split: train
8
+ - path: ["dstc2-rdf/test.jsonl","multiwoz-rdf/test.jsonl","sfxdial-rdf/test.jsonl"]
9
+ split: test
10
+ - path: ["dstc2-rdf/validation.jsonl","multiwoz-rdf/validation.jsonl","sfxdial-rdf/validation.jsonl"]
11
+ split: validation
12
+ - config_name: bundle-simulated
13
+ description: Merge of all rdf simulated datasets
14
+ data_files:
15
+ - path: ["camrest-sim/train.jsonl","multiwoz-sim/train.jsonl"]
16
+ split: train
17
+ - path: ["camrest-sim/test.jsonl","multiwoz-rdf/multiwoz-sim.jsonl"]
18
+ split: test
19
+ - path: ["camrest-sim/validation.jsonl","multiwoz-sim/validation.jsonl"]
20
+ split: validation
21
+ - config_name: dstc2
22
+ description: DSTC2 converted to rdf format
23
+ data_files:
24
+ - path: "dstc2-rdf/train.jsonl"
25
+ split: train
26
+ - path: "dstc2-rdf/test.jsonl"
27
+ split: test
28
+ - path: "dstc2-rdf/validation.jsonl"
29
+ split: validation
30
+ - config_name: sfxdial
31
+ description: Sfxdial converted to rdf format
32
+ data_files:
33
+ - path: "sfxdial-rdf/train.jsonl"
34
+ split: train
35
+ - path: "sfxdial-rdf/test.jsonl"
36
+ split: test
37
+ - path: "sfxdial-rdf/validation.jsonl"
38
+ split: validation
39
+ - config_name: multiwoz
40
+ description: MultiWoz converted to rdf format
41
+ data_files:
42
+ - path: "multiwoz-rdf/train.jsonl"
43
+ split: train
44
+ - path: "multiwoz-rdf/test.jsonl"
45
+ split: test
46
+ - path: "multiwoz-rdf/validation.jsonl"
47
+ split: validation
48
+ - config_name: camrest-sim
49
+ description: Synthetic dialogs on the Cambridge restaurant search domain
50
+ data_files:
51
+ - path: "camrest-sim/train.jsonl"
52
+ split: train
53
+ - path: "camrest-sim/test.jsonl"
54
+ split: test
55
+ - path: "camrest-sim/validation.jsonl"
56
+ split: validation
57
+ - config_name: multiwoz-sim
58
+ description: Synthetic dialogs on the Multiwoz domains
59
+ data_files:
60
+ - path: "multiwoz-sim/train.jsonl"
61
+ split: train
62
+ - path: "multiwoz-sim/test.jsonl"
63
+ split: test
64
+ - path: "multiwoz-sim/validation.jsonl"
65
+ split: validation
66
+ tags:
67
+ - dialogue
68
+ - rdf
69
+ - dst
70
+ task_categories:
71
+ - conversational
72
+ - text-generation
73
+ - text2text-generation
74
+ task_ids:
75
+ - dialogue-state-tracking
76
+ - natural-language-generation
77
+ license:
78
+ - GPL 3.0
79
+ - MIT
80
+ - 'Attribution 2.0 UK: England & Wales'
81
+ - CC BY-NC-SA 4.0
82
+ packages:
83
+ - python-gitlab
84
+ language:
85
+ - fr
86
+ ---
87
+
88
+ # Dataset Card for rdfdial
89
+
90
+
91
+ ## Table of Contents
92
+ - [Table of Contents](#table-of-contents)
93
+ - [Dataset Description](#dataset-description)
94
+ - [Dataset Summary](#dataset-summary)
95
+ - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
96
+ - [Languages](#languages)
97
+ - [Dataset Structure](#dataset-structure)
98
+ - [Data Instances](#data-instances)
99
+ - [Data Fields](#data-fields)
100
+ - [Data Splits](#data-splits)
101
+ - [Dataset Creation](#dataset-creation)
102
+ - [Curation Rationale](#curation-rationale)
103
+ - [Source Data](#source-data)
104
+ - [Annotations](#annotations)
105
+ - [Personal and Sensitive Information](#personal-and-sensitive-information)
106
+ - [Considerations for Using the Data](#considerations-for-using-the-data)
107
+ - [Social Impact of Dataset](#social-impact-of-dataset)
108
+ - [Discussion of Biases](#discussion-of-biases)
109
+ - [Other Known Limitations](#other-known-limitations)
110
+ - [Additional Information](#additional-information)
111
+ - [Dataset Curators](#dataset-curators)
112
+ - [Licensing Information](#licensing-information)
113
+ - [Citation Information](#citation-information)
114
+ - [Contributions](#contributions)
115
+
116
+ ## Dataset Description
117
+
118
+ - **Homepage:** https://huggingface.co/Orange
119
+ - **Repository:** https://huggingface.co/Orange/rdfdial
120
+ - **Paper:** N/A
121
+ - **Leaderboard:** N/A
122
+ - **Point of Contact:** Morgan VEYRET; Lina Maria ROJAS BARAHONA
123
+
124
+ ### Dataset Summary
125
+
126
+ This dataset provides dialogues annotated in dialogue acts and dialogue
127
+ state in and RDF based formalism.
128
+
129
+ There is a conversion of `sfxdial`, `dstc2` and `multiwoz2.3` datasets
130
+ as well as two fully synthetic datasets created from simulated conversations:
131
+ `camrest-sim` and `multiwoz-sim`.
132
+
133
+ Original dataset before conversion are available here:
134
+ - DSTC2: https://github.com/matthen/dstc
135
+ - Multiwoz 2.3: https://github.com/thu-coai/ConvLab-2/tree/master/data/multiwoz2.3
136
+ - SfxDial: https://www.repository.cam.ac.uk/items/62011578-23d4-4355-8878-5a150fb72b43
137
+
138
+ ### Supported Tasks and Leaderboards
139
+
140
+ This dataset was used for the following tasks:
141
+ - Natural Language Generation
142
+ - Dialogue State Tracking
143
+
144
+ ### Languages
145
+
146
+ This dataset includes the following languages:
147
+ - English
148
+
149
+ ## Dataset Structure
150
+
151
+ ### Data Instances
152
+
153
+
154
+ For all datasets, each item has this schema:
155
+
156
+ ```python
157
+ {
158
+ "dialogue_id": "string", # dialog identifier
159
+ "turns": [{ # list of dialog turns
160
+ "id": "int8", # dialog turn index in the conversation
161
+ "speaker": "string", # speaker identifier ('user' or 'system')
162
+ "text": "string", # speaker utterance
163
+ "rdf-acts": ["string"], # string representation of dialog acts
164
+ }],
165
+ "states": [{ # dialog states for each turn
166
+ "id": "int8",
167
+ "multi_relations": "bool", # are multiple instances of relations allowed ?
168
+ "triples": [["string"]], # triples representing the state
169
+ "turn_ids": ["int8"], # ids of turns contributing to this state
170
+ }],
171
+ }
172
+ ```
173
+
174
+ ### Data Fields
175
+
176
+ For each dataset item, the following fields are provided:
177
+ - `dialogue_id`: unique dialogue identifier
178
+ - `turns`: list of speech turns, each turn contains the following fields:
179
+ - `id`: turn index in the dialogue
180
+ - `speaker`: identifier for the speaker (`user` or `system`)
181
+ - `text`: turn utterance
182
+ - `rdf-acts`: list of dialogue acts using string representation of rdf formalism
183
+ each act has the form: `act(triple;...)` where `triple` is formatted as
184
+ `(subject,predicate,object)`
185
+ - `states`: list of states for the dialogue, each entry contains the following fields:
186
+ - `id`: state index in the dialogue
187
+ - `multi_relations`: boolean indicating if multiple instances of the same predicate are
188
+ allowed or not
189
+ - `triples`: list of triples representing the graph state, each triple is a list of 3 string like
190
+ `[subject,predicate,object]`
191
+ - `turn_ids`: list of turn ids that contributed to this state
192
+
193
+
194
+ ### Data Splits
195
+
196
+ For each dataset, splits were generated randomly in the following proportions:
197
+ - *train*: 80%
198
+ - *validation*: 16%
199
+ - *test*: 4%
200
+
201
+ ## Dataset Creation
202
+
203
+ ### Curation Rationale
204
+
205
+ This dataset has been created to work with graph base dialog state representation using
206
+ generative models (T5 family).
207
+
208
+ ### Source Data
209
+
210
+ #### Initial Data Collection and Normalization
211
+
212
+ - *Converted datasets*:
213
+ - DSTC2: https://github.com/matthen/dstc
214
+ - Multiwoz 2.3: https://github.com/thu-coai/ConvLab-2
215
+ - SfxDial: https://www.repository.cam.ac.uk/handle/1810/251304
216
+ - *Synthetic datasets*: rule-based simulations
217
+
218
+ #### Who are the source language producers?
219
+
220
+ - *Converted datasets*: see original datasets documentation
221
+ - *Synthetic datasets*: conversations were generated using an agenda-based user simulator and
222
+ a rule based agent working directly with dialogue acts. These conversations were then augmented
223
+ with natural language user/system utterances. Natural language generation was done using
224
+ a T5-base model fine-tuned on the converted datasets.
225
+
226
+ ### Annotations
227
+
228
+ #### Annotation process
229
+
230
+ - *Converted datasets*: rule-based conversion of the user/system dialogue acts from slot-value
231
+ to RDF based format. The dialogue state is created automatically using another rule based
232
+ tracked working with triples. Some conversations could not be converted automatically and/or
233
+ contained wrong/confusing annotations and were removed from the dataset compared to the
234
+ original ones.
235
+ - *Synthetic datasets*: simulation work at the annotation level and the dataset was augmented
236
+ to include natural language information.
237
+
238
+ #### Who are the annotators?
239
+
240
+ All annotations were generated automatically.
241
+
242
+ For dialogue acts:
243
+ - converted data: rules were applied to convert slot-value based dialogue acts
244
+ to rdf-based ones
245
+ - synthetic data: rdf-based dialogue acts were directly generated by the dialogue simulation.
246
+
247
+ For dialogue states, a rule based system was using taking rdf-based dialogue acts as
248
+ its inputs.
249
+
250
+ ### Personal and Sensitive Information
251
+
252
+ This dataset does not contains any personal or sensitive information.
253
+
254
+ ## Considerations for Using the Data
255
+
256
+ ### Social Impact of Dataset
257
+
258
+ [More Information Needed]
259
+
260
+ ### Discussion of Biases
261
+
262
+ [More Information Needed]
263
+
264
+ ### Other Known Limitations
265
+
266
+ [More Information Needed]
267
+
268
+ ## Additional Information
269
+
270
+ ### Dataset Curators
271
+
272
+ [More Information Needed]
273
+
274
+ ### Licensing Information
275
+
276
+ Converted datasets follow their original licenses:
277
+ - DSTC2: [GPL 3.0](https://github.com/matthen/dstc/blob/master/LICENSE)
278
+ - Multiwoz 2.3: [Apache 2.0](https://github.com/thu-coai/ConvLab-2/blob/master/LICENSE)
279
+ - SfxDial: [Attribution 2.0 UK: England & Wales](https://creativecommons.org/licenses/by/2.0/uk/)
280
+
281
+ Simulated conversation are provided with the following licenses:
282
+ - camrest-sim: [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode)
283
+ - multiwoz-sim: [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode)
284
+
285
+ ### Citation Information
286
+
287
+ [More Information Needed]
288
+
289
+ ### Contributions
290
+
291
+ - Morgan Veyret
camrest-sim-rdf/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a93e309708952b7027cf961d69e9936796267b8b5cc08f28c16d06b9cc4f1f7d
3
+ size 176968
camrest-sim-rdf/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:431ccd8e69da5c2e9c2de08cbde7dd06cb8cd3afa0d79bb27b27c13d90c61218
3
+ size 3621884
camrest-sim-rdf/validation.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d495ff20c5a2d6508aa757aaa0531380093718c611fe6c941c9ddcc13411b1bb
3
+ size 723831
dstc2-rdf/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3007b9a96156a86e7843c2bb27b8804ef7c3a49c32564d6275e639c81ce0ce9a
3
+ size 7944261
dstc2-rdf/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d89ba74534fbf3adc6948c94b42d909c7f999e04e9ba04fd60f36035d415475e
3
+ size 9753439
dstc2-rdf/validation.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38716a0a3da5ea9fe5f771a51b054db80f1498674ccb2ff7e9b89a6e3bed542a
3
+ size 2337938
multiwoz-rdf/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fb803cdbe69d75fbc4b29adb86f3dc4f9bd92783700477b823f65ffee51a7fa
3
+ size 2549543
multiwoz-rdf/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:569d694bebfaf9a1d3815d0c3cf2eac817096bff6c2d4782a751958ab58d3e86
3
+ size 51078387
multiwoz-rdf/validation.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebac00b3c42062b9bd86e72b19eb2ea0788f3ffdb380e8e662c818596e73e98e
3
+ size 10334087
multiwoz-sim-rdf/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e07f19ed8d65cbce4198784045fed215a9a641b06bf4c7a205494ee50a305dcb
3
+ size 157161
multiwoz-sim-rdf/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2bbd2b865cc3e28081b6894c879dd7ef822638ca86da4960f9aaa416c278b905
3
+ size 3135842
multiwoz-sim-rdf/validation.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93959a3a7a8c5d331071dde34b92fa7ed97c3c13d96c28fc1fbf3d116f278225
3
+ size 658219
sfxdial-rdf/test.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d569a7d235f2801d0c8862828fca49cb23bb8ff41474748645ee93a8b4dffc29
3
+ size 378468
sfxdial-rdf/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3cf4ba10e4e2ba4fcce8cab79e0301008664295f2b3a9beab3959c97e45faa33
3
+ size 7435171
sfxdial-rdf/validation.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e006c85e884b8c77170ea1bba76144c2234855dac1a9a8791eb3c13d78d2ad79
3
+ size 1522698