Jack Rueter - Academia.edu (original) (raw)
Papers by Jack Rueter
Zenodo (CERN European Organization for Nuclear Research), 2023
Two-level morphophonology and lexc morphological descriptions of Skolt Sami as coded on the Giell... more Two-level morphophonology and lexc morphological descriptions of Skolt Sami as coded on the Giella Infrastructure at the Arctic University of Norway, in Tromsø. Funded by scholarship from Kone Foundation
Morphological models for generation, lemmatization and analysis in 22 languages. The models are t... more Morphological models for generation, lemmatization and analysis in 22 languages. The models are trained in OpenNMT-py https://github.com/OpenNMT/OpenNMT-py. Feed one word at a time, split into characters (kissa -> k i s s a) Supported languages: German (deu), Kven (fkv), Komi-Zyrian (kpv), Mokhsa (mdf), Mansi (mns), Erzya (myv), Norwegian Bokmål (nob), Russian (rus), South Sami (sma), Lule Sami (smj), Skolt Sami (sms), Võro (vro), Finnish (fin), Komi-Permyak (koi), Latvian (lav), Eastern Mari (mhr), Western Mari (mrj), Namonuito (nmt), Olonets-Karelian (olo), Pite Sami (sje), Northern Sami (sme), Inari Sami (smn) and Udmurt (udm)
This initial pre-release contains three maps representing the locales where fieldwork was done co... more This initial pre-release contains three maps representing the locales where fieldwork was done collecting Erzya and Moksha language materials for and by Heikki Paasonen 1891–1912.
The purpose of this article is to outline morphological facts about the two literary languages Er... more The purpose of this article is to outline morphological facts about the two literary languages Erzya and Moksha, which can be used for estimating the distinctive character of these individual language forms. Whereas earlier morphological evaluations of the linguistic distance between Erzya and Moksha have placed them in the area of 90% cohesion, this one does not. This study evaluates the languages on the basis of non-ambiguity, parallel sets of ambiguity and divergent ambiguity. Non-ambiguity is found in combinatory function to morphological formant alignment, e.g. молян go+V+Ind+Prs+ScSg1. Parallel sets of ambiguity is found in combinatory-function set to morphological formant alignment where both languages share the same sets of ambiguous readings, e.g. саизь v s сявозь take+V+Ind+ScPl3+OcSg3, ScPl3+OcPl3. Divergent ambiguity is found in forms with non- symmetric alignments of combinatory functions, e.g. саинек take+V+Ind+Prt1+ScPl1, +Prt1+ScPl1+OcSg3, +Prt1+ScPl1+OcPl3 vs сявоме take+V+Ind+Prt1+ScPl1, сявоськ take+V+Ind+Prt1+ScPl1+OcSg3, +Prt1+ScPl1+OcPl3. This morphological evaluation will establish the preparatory work in syntactic disambiguation necessary for facilitating Erzya↔Moksha machine translation, whereas machine translation will enhance the usage of mutual language resources. Results show that the Erzya and Moksha languages, in the absence of loan words from the 20 th century, share less than 50% of their vocabularies, 63% of their regular nominal declensions and 48% of their regular finite conjugations.Peer reviewe
Routledge eBooks, Feb 20, 2023
Zenodo (CERN European Organization for Nuclear Research), Feb 22, 2023
Zenodo (CERN European Organization for Nuclear Research), 2023
Two-level morphophonology and lexc morphological descriptions of Skolt Sami as coded on the Giell... more Two-level morphophonology and lexc morphological descriptions of Skolt Sami as coded on the Giella Infrastructure at the Arctic University of Norway, in Tromsø. Funded by scholarship from Kone Foundation
Morphological models for generation, lemmatization and analysis in 22 languages. The models are t... more Morphological models for generation, lemmatization and analysis in 22 languages. The models are trained in OpenNMT-py https://github.com/OpenNMT/OpenNMT-py. Feed one word at a time, split into characters (kissa -> k i s s a) Supported languages: German (deu), Kven (fkv), Komi-Zyrian (kpv), Mokhsa (mdf), Mansi (mns), Erzya (myv), Norwegian Bokmål (nob), Russian (rus), South Sami (sma), Lule Sami (smj), Skolt Sami (sms), Võro (vro), Finnish (fin), Komi-Permyak (koi), Latvian (lav), Eastern Mari (mhr), Western Mari (mrj), Namonuito (nmt), Olonets-Karelian (olo), Pite Sami (sje), Northern Sami (sme), Inari Sami (smn) and Udmurt (udm)
This initial pre-release contains three maps representing the locales where fieldwork was done co... more This initial pre-release contains three maps representing the locales where fieldwork was done collecting Erzya and Moksha language materials for and by Heikki Paasonen 1891–1912.
The purpose of this article is to outline morphological facts about the two literary languages Er... more The purpose of this article is to outline morphological facts about the two literary languages Erzya and Moksha, which can be used for estimating the distinctive character of these individual language forms. Whereas earlier morphological evaluations of the linguistic distance between Erzya and Moksha have placed them in the area of 90% cohesion, this one does not. This study evaluates the languages on the basis of non-ambiguity, parallel sets of ambiguity and divergent ambiguity. Non-ambiguity is found in combinatory function to morphological formant alignment, e.g. молян go+V+Ind+Prs+ScSg1. Parallel sets of ambiguity is found in combinatory-function set to morphological formant alignment where both languages share the same sets of ambiguous readings, e.g. саизь v s сявозь take+V+Ind+ScPl3+OcSg3, ScPl3+OcPl3. Divergent ambiguity is found in forms with non- symmetric alignments of combinatory functions, e.g. саинек take+V+Ind+Prt1+ScPl1, +Prt1+ScPl1+OcSg3, +Prt1+ScPl1+OcPl3 vs сявоме take+V+Ind+Prt1+ScPl1, сявоськ take+V+Ind+Prt1+ScPl1+OcSg3, +Prt1+ScPl1+OcPl3. This morphological evaluation will establish the preparatory work in syntactic disambiguation necessary for facilitating Erzya↔Moksha machine translation, whereas machine translation will enhance the usage of mutual language resources. Results show that the Erzya and Moksha languages, in the absence of loan words from the 20 th century, share less than 50% of their vocabularies, 63% of their regular nominal declensions and 48% of their regular finite conjugations.Peer reviewe
Routledge eBooks, Feb 20, 2023
Zenodo (CERN European Organization for Nuclear Research), Feb 22, 2023
Proceedings of the 3rd Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature, 2019
This paper presents multiple methods for normalizing the most deviant and infrequent historical s... more This paper presents multiple methods for normalizing the most deviant and infrequent historical spellings in a corpus consisting of personal correspondence from the 15th to the 19th century. The methods include machine translation (neural and statistical), edit distance and rule-based FST. Different normalization methods are compared and evaluated. All of the methods have their own strengths in word normalization. This calls for finding ways of combining the results from these methods to leverage their individual strengths.