update to use Split.TRAIN

README.md

@@ -64,7 +64,7 @@ This is a stance detection dataset in the Zulu language. The data is translated
 
 Our paper aims at utilizing this progress made for English to transfers that knowledge into other languages, which is a non-trivial task due to the domain gap between English and the target languages. We propose a black-box non-intrusive method that utilizes techniques from Domain Adaptation to reduce the domain gap, without requiring any human expertise in the target language, by leveraging low-quality data in both a supervised and unsupervised manner. This allows us to rapidly achieve similar results for stance detection for the Zulu language, the target language in this work, as are found for English. A natively-translated dataset is used for evaluation of domain transfer.
 
-This dataset only defines a `TEST` partition.
+This dataset defines a validation partition, as per the scenario in the paper.
 
 ### Supported Tasks and Leaderboards
 

dataset_infos.json

@@ -1 +1 @@
-{"zulu-stance": {"description": "Misinformation has become a major concern in recent last years given its \nspread across our information sources. In the past years, many NLP tasks have\nbeen introduced in this area, with some systems reaching good results on \nEnglish language datasets. Existing AI based approaches for fighting \nmisinformation in literature suggest automatic stance detection as an integral\nfirst step to success. Our paper aims at utilizing this progress made for\nEnglish to transfers that knowledge into other languages, which is a \nnon-trivial task due to the domain gap between English and the target \nlanguages. We propose a black-box non-intrusive method that utilizes techniques\nfrom Domain Adaptation to reduce the domain gap, without requiring any human\nexpertise in the target language, by leveraging low-quality data in both a\nsupervised and unsupervised manner. This allows us to rapidly achieve similar\nresults for stance detection for the Zulu language, the target language in\nthis work, as are found for English. We also provide a stance detection dataset\nin the Zulu language. \n", "citation": "@inproceedings{dlamini_zulu_stance,\n  title={Bridging the Domain Gap for Stance Detection for the Zulu language},\n  author={Dlamini, Gcinizwe and Bekkouch, Imad Eddine Ibrahim and Khan, Adil and Derczynski, Leon},\n  booktitle={Proceedings of IEEE IntelliSys},\n  year={2022}\n}\n", "homepage": "https://arxiv.org/abs/2205.03153", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "stance": {"num_classes": 3, "names": ["FAVOR", "AGAINST", "NONE"], "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "zulu_stance", "config_name": "zulu-stance", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 191244, "num_examples": 1343, "dataset_name": "zulu_stance"}}, "download_checksums": {"ZUstance.json": {"num_bytes": 217638, "checksum": "dcf109810d5c511ea0bb02e1fc0b6a5278a02b1d581e57840f33bed1b2b1b0c7"}}, "download_size": 217638, "post_processing_size": null, "dataset_size": 191244, "size_in_bytes": 408882}}
+{"zulu-stance": {"description": "This is a stance detection dataset in the Zulu language. The data is translated to Zulu by Zulu native speakers, from English source texts.\n\nMisinformation has become a major concern in recent last years given its \nspread across our information sources. In the past years, many NLP tasks have\nbeen introduced in this area, with some systems reaching good results on \nEnglish language datasets. Existing AI based approaches for fighting \nmisinformation in literature suggest automatic stance detection as an integral\nfirst step to success. Our paper aims at utilizing this progress made for\nEnglish to transfers that knowledge into other languages, which is a \nnon-trivial task due to the domain gap between English and the target \nlanguages. We propose a black-box non-intrusive method that utilizes techniques\nfrom Domain Adaptation to reduce the domain gap, without requiring any human\nexpertise in the target language, by leveraging low-quality data in both a\nsupervised and unsupervised manner. This allows us to rapidly achieve similar\nresults for stance detection for the Zulu language, the target language in\nthis work, as are found for English. We also provide a stance detection dataset\nin the Zulu language. \n", "citation": "@inproceedings{dlamini_zulu_stance,\n  title={Bridging the Domain Gap for Stance Detection for the Zulu language},\n  author={Dlamini, Gcinizwe and Bekkouch, Imad Eddine Ibrahim and Khan, Adil and Derczynski, Leon},\n  booktitle={Proceedings of IEEE IntelliSys},\n  year={2022}\n}\n", "homepage": "https://arxiv.org/abs/2205.03153", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "stance": {"num_classes": 3, "names": ["FAVOR", "AGAINST", "NONE"], "id": null, "_type": "ClassLabel"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "zulu_stance", "config_name": "zulu-stance", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 191244, "num_examples": 1343, "dataset_name": "zulu_stance"}}, "download_checksums": {"ZUstance.json": {"num_bytes": 217638, "checksum": "dcf109810d5c511ea0bb02e1fc0b6a5278a02b1d581e57840f33bed1b2b1b0c7"}}, "download_size": 217638, "post_processing_size": null, "dataset_size": 191244, "size_in_bytes": 408882}}

zulu_stance.py

@@ -35,6 +35,8 @@ _CITATION = """\
 """
 
 _DESCRIPTION = """\
+This is a stance detection dataset in the Zulu language. The data is translated to Zulu by Zulu native speakers, from English source texts.
+
 Misinformation has become a major concern in recent last years given its 
 spread across our information sources. In the past years, many NLP tasks have
 been introduced in this area, with some systems reaching good results on