Converting a comprehensive lexical database into a computational
            model: The case of East Cree verb inflection

Arppe, Antti; Junker, Marie-Odile; Torkornoo, Delasie

doi:10.18653/v1/w17-0108

Cited by 4 publications

(4 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Dunham (2014) created two morphological parsers which use Finite State Transducers (FST) to model the complex morphophonology of Blackfoot; it is possible that these could be modified for our use. There are other examples of FST parsers for Algonquian languages: for Cree see Arppe et al (2017) and Harrigan et al (2017); for Arapaho see Kazeminejad et al (2017). The current Plains Cree FST parser is available on github (https:// giell alt.…”

Section: Derivational and Inflectional Morphologymentioning

confidence: 99%

Blackfoot Words: a database of Blackfoot lexical forms

Weber

Brown

Celli

et al. 2023

Lang Resources & Evaluation

View full text Add to dashboard Cite

This paper describes the structure and creation of Blackfoot Words, a new relational database of lexical forms (inflected words, stems, and morphemes) in Blackfoot (Algonquian; ISO 639-3: bla). To date, we have digitized 63,493 individual lexical forms from 30 sources, representing all four major dialects, and spanning the years 1743–2017. Version 1.1 of the database includes lexical forms from nine of these sources. This project has two aims. The first is to digitize and provide access to the lexical data in these sources, many of which are difficult to access and discover. The second is to organize the data so that connections can be made between instances of the “same” lexical form across all sources, despite variation across sources in the dialect recorded, orthographic conventions, and the depth of morpheme analysis. The database structure was developed in response to these aims. The database comprises five tables: Sources, Words, Stems, Morphemes, and Lemmas. The Sources table contains bibliographic information and commentary on the sources. The Words table contains inflected words in the source orthography. Each word is broken down into stems and morphemes which are entered into the Stems and Morphemes tables in the source orthography. The Lemmas table contains abstract versions of each stem or morpheme in a standardized orthography. Instances of the same stem or morpheme are linked to a common lemma. We expect that the database will support projects by the language community and other researchers.

show abstract

Section: Derivational and Inflectional Morphologymentioning

confidence: 99%

Blackfoot Words: a database of Blackfoot lexical forms

Weber

Brown

Celli

et al. 2023

Lang Resources & Evaluation

View full text Add to dashboard Cite

show abstract

“…More recently, an Innu-Aimun morphological segmenter based on deep-learning was proposed [35]. Fundamental resources like morphological models for segmentation exist for related languages, namely East Cree [1] and Plains Cree [33] [15]. Such resources could potentially be adapted to Innu-Aimun.…”

Section: Related Workmentioning

confidence: 99%

Lost in Innu-Aimun Translation - Re-dening Neural Machine Translation for Indigenous Interpreters and Translators Needs

Cadotte,

Thernish,

Sadat

2023

Proceedings of the International Conference on Human-Informed Translation and Interpreting Technology 2023

View full text Add to dashboard Cite

Innu-Aimun, one of the most spoken Indigenous languages in Canada, faces signicant transmission challenges. Although there is a notable body of Innu-Aimun literature, there is generally not enough documentation written in Innu-Aimun for daily use, just as there are not enough translators and interpreters for the language. We present here collaborative work between Innu-Aimun translators and researchers in computational linguistics to develop translation assistance tools, with the aim of helping language revitalization and preservation. We detail our common position on how should technological assistance tools be developed for Innu-Aimun, which emphasizes the importance of involving Innu translators throughout the entire process and making sure to address language-specic needs. This position is elaborated from joining our respective perspectives (researchers and Innu community member) and expertise (computational linguistics and Innu-Aimun translation). In this spirit, we present preliminary results for the rst ongoing steps towards building a rst Innu-Aimun -French Neural Machine Translation model. We focus on our participatory process to create aligned parallel corpora and present rst results and analyses.

show abstract

“…The proposed approach builds on the established practice of applying finite state transducers for modelling the phonological and morphological systems of natural languages (Beesley and Karttunen, 2003). In recent years, researchers have demonstrated the suitability of finite state models for a variety of morphologically complex, low-resource languages including Cree, Haida, Kunwinjku, Odawa, Tsuut'ina, and Yupik (Snoek et al, 2014;Harrigan et al, 2017;Arppe et al, 2017a;Arppe et al, 2017b;Bowers et al, 2017;Chen and Schwartz, 2018;Lachler et al, 2018;Lane and Bird, 2019).…”

Section: Finite State Analysis For Morphologically Complex Languagesmentioning

confidence: 99%

Interactive Word Completion for Morphologically Complex Languages

Lane

Bird

2020

Proceedings of the 28th International Conference on Computational Linguistics

View full text Add to dashboard Cite

Text input technologies for low-resource languages support literacy, content authoring, and language learning. However, tasks such as word completion pose a challenge for morphologically complex languages thanks to the combinatorial explosion of possible words. We have developed a method for morphologically-aware text input in Kunwinjku, a polysynthetic language of northern Australia. We modify an existing finite state recognizer to map input morph prefixes to morph completions, respecting the morphosyntax and morphophonology of the language. We demonstrate the portability of the method by applying it to Turkish. We show that the space of proximal morph completions is many orders of magnitude smaller than the space of full word completions for Kunwinjku. We provide a visualization of the morph completion space to enable the text completion parameters to be fine-tuned. Finally, we report on a web services deployment, along with a web interface which helps users enter morphologically complex words and which retrieves corresponding entries from the lexicon.

show abstract

Converting a comprehensive lexical database into a computational model: The case of East Cree verb inflection

Cited by 4 publications

References 3 publications

Blackfoot Words: a database of Blackfoot lexical forms

Blackfoot Words: a database of Blackfoot lexical forms

Lost in Innu-Aimun Translation - Re-dening Neural Machine Translation for Indigenous Interpreters and Translators Needs

Interactive Word Completion for Morphologically Complex Languages

Contact Info

Product

Resources

About