Datasets:
Size:
10K<n<100K
License:
Pierre Colombo
commited on
Commit
•
23d9f85
1
Parent(s):
2299233
Update documentation card of miam dataset (#4846)
Browse files* Update README.md
* Fix dataset card
Co-authored-by: Albert Villanova del Moral <8515462+albertvillanova@users.noreply.github.com>
Commit from https://github.com/huggingface/datasets/commit/5caced4d733d2b49f3bd2572512b7c15cb22d865
README.md
CHANGED
@@ -240,9 +240,9 @@ For the `vm2` configuration, the different fields are:
|
|
240 |
|
241 |
## Additional Information
|
242 |
|
243 |
-
###
|
244 |
|
245 |
-
Anonymous
|
246 |
|
247 |
### Licensing Information
|
248 |
|
@@ -251,13 +251,24 @@ This work is licensed under a [Creative Commons Attribution-NonCommercial-ShareA
|
|
251 |
### Citation Information
|
252 |
|
253 |
```
|
254 |
-
@
|
255 |
-
|
256 |
-
|
257 |
-
|
258 |
-
|
259 |
-
|
260 |
-
|
261 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
262 |
}
|
263 |
```
|
|
|
|
|
|
|
|
|
|
240 |
|
241 |
## Additional Information
|
242 |
|
243 |
+
### Dataset Curators
|
244 |
|
245 |
+
Anonymous.
|
246 |
|
247 |
### Licensing Information
|
248 |
|
|
|
251 |
### Citation Information
|
252 |
|
253 |
```
|
254 |
+
@inproceedings{colombo-etal-2021-code,
|
255 |
+
title = "Code-switched inspired losses for spoken dialog representations",
|
256 |
+
author = "Colombo, Pierre and
|
257 |
+
Chapuis, Emile and
|
258 |
+
Labeau, Matthieu and
|
259 |
+
Clavel, Chlo{\'e}",
|
260 |
+
booktitle = "Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing",
|
261 |
+
month = nov,
|
262 |
+
year = "2021",
|
263 |
+
address = "Online and Punta Cana, Dominican Republic",
|
264 |
+
publisher = "Association for Computational Linguistics",
|
265 |
+
url = "https://aclanthology.org/2021.emnlp-main.656",
|
266 |
+
doi = "10.18653/v1/2021.emnlp-main.656",
|
267 |
+
pages = "8320--8337",
|
268 |
+
abstract = "Spoken dialogue systems need to be able to handle both multiple languages and multilinguality inside a conversation (\textit{e.g} in case of code-switching). In this work, we introduce new pretraining losses tailored to learn generic multilingual spoken dialogue representations. The goal of these losses is to expose the model to code-switched language. In order to scale up training, we automatically build a pretraining corpus composed of multilingual conversations in five different languages (French, Italian, English, German and Spanish) from OpenSubtitles, a huge multilingual corpus composed of 24.3G tokens. We test the generic representations on MIAM, a new benchmark composed of five dialogue act corpora on the same aforementioned languages as well as on two novel multilingual tasks (\textit{i.e} multilingual mask utterance retrieval and multilingual inconsistency identification). Our experiments show that our new losses achieve a better performance in both monolingual and multilingual settings.",
|
269 |
}
|
270 |
```
|
271 |
+
|
272 |
+
### Contributions
|
273 |
+
|
274 |
+
Thanks to [@eusip](https://github.com/eusip) and [@PierreColombo](https://github.com/PierreColombo) for adding this dataset.
|