Search

Start Over Language Slovak Publisher Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL) Rights Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)

1. DaMuEL 1.0: A Large Multilingual Dataset for Entity Linking

Creator:
Kubeša, David and Straka, Milan
Publisher:
Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)
Type:
text and corpus
Subject:
entity linking, NEL, NER, dataset, and knowledge base
Language:
Afrikaans, Arabic, Armenian, Basque, Belarusian, Bulgarian, Catalan, Chinese, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, Galician, German, Hebrew, Hindi, Hungarian, Indonesian, Irish, Italian, Japanese, Korean, Latin, Latvian, Lithuanian, Maltese, Marathi, Modern Greek (1453-), Northern Sami, Norwegian Nynorsk, Persian, Polish, Portuguese, Romanian, Russian, Scottish Gaelic, Serbian, Slovak, Slovenian, Spanish, Swedish, Tamil, Telugu, Uighur, Ukrainian, Urdu, Vietnamese, and Wolof
Description:
We present DaMuEL, a large Multilingual Dataset for Entity Linking containing data in 53 languages. DaMuEL consists of two components: a knowledge base that contains language-agnostic information about entities, including their claims from Wikidata and named entity types (PER, ORG, LOC, EVENT, BRAND, WORK_OF_ART, MANUFACTURED); and Wikipedia texts with entity mentions linked to the knowledge base, along with language-specific text from Wikidata such as labels, aliases, and descriptions, stored separately for each language. The Wikidata QID is used as a persistent, language-agnostic identifier, enabling the combination of the knowledge base with language-specific texts and information for each entity. Wikipedia documents deliberately annotate only a single mention for every entity present; we further automatically detect all mentions of named entities linked from each document. The dataset contains 27.9M named entities in the knowledge base and 12.3G tokens from Wikipedia texts. The dataset is published under the CC BY-SA licence.
Rights:
Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), http://creativecommons.org/licenses/by-sa/4.0/, and PUB

2. Deltacorpus

Creator:
Mareček, David, Yu, Zhiwei, Zeman, Daniel, and Žabokrtský, Zdeněk
Publisher:
Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)
Type:
text and corpus
Subject:
part of speech, tagging, semi-supervised, and cross-language
Language:
Belarusian, Bosnian, Bulgarian, Czech, Serbo-Croatian, Croatian, Upper Sorbian, Macedonian, Polish, Russian, Slovak, Slovenian, Serbian, Ukrainian, Latvian, Lithuanian, Afrikaans, Danish, German, English, Faroese, Western Frisian, Swiss German, Icelandic, Limburgan, Luxembourgish, Low German, Dutch, Norwegian Nynorsk, Norwegian, Scots, Swedish, Yiddish, Aragonese, Asturian, Catalan, French, Galician, Haitian, Italian, Latin, Lombard, Neapolitan, Piemontese, Portuguese, Romanian, Spanish, Venetian, Walloon, Breton, Welsh, Scottish Gaelic, Irish, Modern Greek (1453-), Armenian, Albanian, Dimli (individual language), Persian, Gilaki, Kurdish, Tajik, Bengali, Bishnupriya, Gujarati, Fiji Hindi, Hindi, Marathi, Nepali (macrolanguage), Urdu, Amharic, Arabic, Egyptian Arabic, Hebrew, Estonian, Finnish, Hungarian, Basque, Georgian, Chuvash, Azerbaijani, Turkish, Uzbek, Kazakh, Tatar, Yakut, Korean, Mongolian, Telugu, Kannada, Malayalam, Tamil, Newari, Vietnamese, Indonesian, Javanese, Malagasy, Maori, Malay (macrolanguage), Pampanga, Sundanese, Tagalog, Waray (Philippines), Swahili (macrolanguage), Esperanto, Ido, Interlingua (International Auxiliary Language Association), and Volapük
Description:
Texts in 107 languages from the W2C corpus (http://hdl.handle.net/11858/00-097C-0000-0022-6133-9), first 1,000,000 tokens per language, tagged by the delexicalized tagger described in Yu et al. (2016, LREC, Portorož, Slovenia).
Rights:
Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), http://creativecommons.org/licenses/by-sa/4.0/, and PUB

3. Deltacorpus 1.1

Creator:
Mareček, David, Yu, Zhiwei, Zeman, Daniel, and Žabokrtský, Zdeněk
Publisher:
Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)
Type:
text and corpus
Subject:
part of speech, tagging, semi-supervised, and cross-language
Language:
Belarusian, Bosnian, Bulgarian, Czech, Serbo-Croatian, Croatian, Upper Sorbian, Macedonian, Polish, Russian, Slovak, Slovenian, Serbian, Ukrainian, Latvian, Lithuanian, Afrikaans, Danish, German, English, Faroese, Western Frisian, Swiss German, Icelandic, Limburgan, Luxembourgish, Low German, Dutch, Norwegian Nynorsk, Norwegian, Scots, Swedish, Yiddish, Aragonese, Asturian, Catalan, French, Galician, Haitian, Italian, Latin, Lombard, Neapolitan, Piemontese, Portuguese, Romanian, Spanish, Venetian, Walloon, Breton, Welsh, Scottish Gaelic, Irish, Modern Greek (1453-), Armenian, Albanian, Dimli (individual language), Persian, Gilaki, Kurdish, Tajik, Bengali, Bishnupriya, Gujarati, Fiji Hindi, Hindi, Marathi, Nepali (macrolanguage), Urdu, Amharic, Arabic, Egyptian Arabic, Hebrew, Estonian, Finnish, Hungarian, Basque, Georgian, Chuvash, Azerbaijani, Turkish, Uzbek, Kazakh, Tatar, Yakut, Korean, Mongolian, Telugu, Kannada, Malayalam, Tamil, Newari, Vietnamese, Indonesian, Javanese, Malagasy, Maori, Malay (macrolanguage), Pampanga, Sundanese, Tagalog, Waray (Philippines), Swahili (macrolanguage), Esperanto, Ido, Interlingua (International Auxiliary Language Association), and Volapük
Description:
Texts in 107 languages from the W2C corpus (http://hdl.handle.net/11858/00-097C-0000-0022-6133-9), first 1,000,000 tokens per language, tagged by the delexicalized tagger described in Yu et al. (2016, LREC, Portorož, Slovenia). Changes in version 1.1: 1. Universal Dependencies tagset instead of the older and smaller Google Universal POS tagset. 2. SVM classifier trained on Universal Dependencies 1.2 instead of HamleDT 2.0. 3. Balto-Slavic languages, Germanic languages and Romance languages were tagged by classifier trained only on the respective group of languages. Other languages were tagged by a classifier trained on all available languages. The "c7" combination from version 1.0 is no longer used.
Rights:
Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), http://creativecommons.org/licenses/by-sa/4.0/, and PUB

4. UDify Pretrained Model

Creator:
Kondratyuk, Dan and Straka, Milan
Publisher:
Charles University, Faculty of Mathematics and Physics, Institute of Formal and Applied Linguistics (UFAL)
Type:
tool and toolService
Subject:
syntax, dependency parser, and universal dependencies
Language:
Ancient Greek (to 1453), Arabic, Basque, Bulgarian, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Gothic, Modern Greek (1453-), Hebrew, Hindi, Hungarian, Indonesian, Irish, Italian, Japanese, Latin, Norwegian, Church Slavic, Persian, Polish, Portuguese, Romanian, Slovenian, Spanish, Swedish, Tamil, Catalan, Chinese, Galician, Kazakh, Latvian, Russian, Turkish, Coptic, Sanskrit, Slovak, Ukrainian, Uighur, Vietnamese, Belarusian, Korean, Lithuanian, Urdu, Russia Buriat, Northern Kurdish, Northern Sami, Upper Sorbian, Afrikaans, Yue Chinese, Marathi, Serbian, Swedish Sign Language, Telugu, Amharic, Armenian, Breton, Faroese, Komi-Zyrian, Nigerian Pidgin, Old French (842-ca. 1400), Tagalog, Thai, Warlpiri, Yoruba, Akkadian, Bambara, Erzya, and Maltese
Description:
Pretrained model weights for the UDify model, and extracted BERT weights in pytorch-transformers format. Note that these weights slightly differ from those used in the paper.
Rights:
Creative Commons - Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), http://creativecommons.org/licenses/by-sa/4.0/, and PUB