| { | |
| "paper_id": "2022", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T16:22:21.892910Z" | |
| }, | |
| "title": "Deep Learning-Based Morphological Segmentation for Indigenous Languages: A Study Case on Innu-Aimun", | |
| "authors": [ | |
| { | |
| "first": "Ngoc", | |
| "middle": [ | |
| "Tan" | |
| ], | |
| "last": "Le", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Universit\u00e9 du Qu\u00e9bec \u00e0 Montr\u00e9al", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Antoine", | |
| "middle": [], | |
| "last": "Cadotte", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Universit\u00e9 du Qu\u00e9bec \u00e0", | |
| "location": { | |
| "settlement": "Montr\u00e9al" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Mathieu", | |
| "middle": [], | |
| "last": "Boivin", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Universit\u00e9 de Montr\u00e9al", | |
| "location": {} | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Fatiha", | |
| "middle": [], | |
| "last": "Sadat", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Universit\u00e9 du Qu\u00e9bec \u00e0", | |
| "location": { | |
| "settlement": "Montr\u00e9al" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "Recent advances in the field of deep learning have led to a growing interest in the development of NLP approaches for low-resource and endangered languages. Nevertheless, relatively little research, related to NLP, has been conducted on indigenous languages. These languages are considered to be filled with complexities and challenges that make their study incredibly difficult in the NLP and AI fields. This paper focuses on the morphological segmentation of indigenous languages, an extremely challenging task because of polysynthesis, dialectal variations with rich morpho-phonemics, misspellings and resource-limited scenario issues. The proposed approach, towards a morphological segmentation of Innu-Aimun, an extremely low-resource indigenous language of Canada, is based on deep learning. Experiments and evaluations have shown promising results, compared to state-of-the-art rule-based and unsupervised approaches.", | |
| "pdf_parse": { | |
| "paper_id": "2022", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "Recent advances in the field of deep learning have led to a growing interest in the development of NLP approaches for low-resource and endangered languages. Nevertheless, relatively little research, related to NLP, has been conducted on indigenous languages. These languages are considered to be filled with complexities and challenges that make their study incredibly difficult in the NLP and AI fields. This paper focuses on the morphological segmentation of indigenous languages, an extremely challenging task because of polysynthesis, dialectal variations with rich morpho-phonemics, misspellings and resource-limited scenario issues. The proposed approach, towards a morphological segmentation of Innu-Aimun, an extremely low-resource indigenous language of Canada, is based on deep learning. Experiments and evaluations have shown promising results, compared to state-of-the-art rule-based and unsupervised approaches.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Over the past decade, we have observed a successful growth in the deep learning-based approaches in several Natural Language Processing (NLP) applications. This has helped to create NLP tools and applications in resource-rich languages. On the other hand, for low-resource languages, few applications of NLP have been studied for multiple reasons (Mager et al., 2018b) .", | |
| "cite_spans": [ | |
| { | |
| "start": 347, | |
| "end": 368, | |
| "text": "(Mager et al., 2018b)", | |
| "ref_id": "BIBREF24" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "In particular, for indigenous languages, NLP applications have to deal with linguistics challenges such as polysynthesis, diversity of grammatical features of morphology, dialect variation with rich morpho-phonemics, misspellings due to noisy or scarce training data and low resource scenario challenges (Littell et al., 2018; Joanis et al., 2020) . Moreover, morphological segmentation for indigenous polysynthetic languages is especially challenging because these languages have often multiple individual morphemes by word and several meanings per morpheme.", | |
| "cite_spans": [ | |
| { | |
| "start": 304, | |
| "end": 326, | |
| "text": "(Littell et al., 2018;", | |
| "ref_id": "BIBREF21" | |
| }, | |
| { | |
| "start": 327, | |
| "end": 347, | |
| "text": "Joanis et al., 2020)", | |
| "ref_id": "BIBREF13" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "The current research focuses on the morphological segmentation task for indigenous languages, with a case study on Innu-Aimun, also called Montagnais 1 . Innu-Aimun is an Algonquian polysynthetic language spoken by over 10,000 Innu in Labrador and Quebec in Eastern Canada 2 . We choose this indigenous language for this specific NLP task because it has not yet been investigated thus far.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "The main focus consists of how to develop indigenous language technology and linguistic resources, with the aim of helping the indigenous communities in the revitalization and preservation of their languages. Thus, we propose in the current study, a deep learning-based morphological segmentation for Innu-Aimun. Our contribution to the current research is twofold. Firstly, it proposes a deep learning-based word segmenter for indigenous languages. Secondly, it empirically compares the proposed approach, in a case study of Innu-Aimun, with multiple baselines such as Finite-State Transducer, Morfessor, and Adaptor Grammar-based approaches.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Overall, this study aims to serve as a benchmark for developing NLP tools and applications, which will help revitalize and preserve indigenous languages, while taking into account indigenous cultural realities and knowledge.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "The paper is structured as follows: Section 2 highlights morphological analyzers for indigenous languages, with a description of Innu-Aimun. Our proposed approach is described in Section 3. Section 4 presents the experimental results, compared to other state-of-the-art approaches. Section 5 discusses our evaluations, while providing an error analysis. Finally, Section 6 presents the conclusion as well as potential future work.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "Many indigenous languages in Canada, in the Americas and around the world have in common that they are polysynthetic. Most also share a context of extremely low or scarce resource. While morphological segmentation is highly useful-if not unavoidable-for indigenous NLP applications, data and knowledge scarcity make its development very challenging. When there exists no language-specific tool, NLP tasks often make use of unsupervised approaches for segmentation. Byte-pair encoding (BPE) segmentation, introduced by Sennrich et al. 2016, is a common one for Neural Machine Translation. The technique has been used by (Joanis et al., 2020; Le and Sadat, 2020) , for instance, to produce an Inuktitut-English NMT baseline using the Nunavut Hansard corpus.", | |
| "cite_spans": [ | |
| { | |
| "start": 619, | |
| "end": 640, | |
| "text": "(Joanis et al., 2020;", | |
| "ref_id": "BIBREF13" | |
| }, | |
| { | |
| "start": 641, | |
| "end": 660, | |
| "text": "Le and Sadat, 2020)", | |
| "ref_id": "BIBREF20" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Related work 2.1 Morphological segmentation in Indigenous languages", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "In cases where there is a lack of annotated data, rule-based approaches, such as those based on Other proposed approaches are hybrid, adding knowledge or rules to unsupervised methods. Eskander et al. (2019) proposed an approach based on Adaptor Grammars (Johnson et al., 2006) , and applied it to four Uto-Aztecan polysynthetic languages. Pan et al. (2020) combined BPE segmentation and rule-based segmentation for Uyghur, a morphologically rich language.", | |
| "cite_spans": [ | |
| { | |
| "start": 185, | |
| "end": 207, | |
| "text": "Eskander et al. (2019)", | |
| "ref_id": "BIBREF10" | |
| }, | |
| { | |
| "start": 255, | |
| "end": 277, | |
| "text": "(Johnson et al., 2006)", | |
| "ref_id": "BIBREF14" | |
| }, | |
| { | |
| "start": 340, | |
| "end": 357, | |
| "text": "Pan et al. (2020)", | |
| "ref_id": "BIBREF27" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Related work 2.1 Morphological segmentation in Indigenous languages", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "For deep learning-based approach, Kann et al. (2018) used the neural network-based seq2seq models for Mexican polysynthetic languages. Micher (2019) applied a recurrent neural network-based approach to deal with the word segmentation for Inuktitut.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Related work 2.1 Morphological segmentation in Indigenous languages", | |
| "sec_num": "2" | |
| }, | |
| { | |
| "text": "Innu-Aimun is the language of the Innu, an indigenous people formerly known as the Montagnais (Mollen, 2006) . This language is found in the Quebec and Labrador provinces of Canada, in a dozen communities (Baraby et al., 2017) . It is a polysynthetic indigenous language, a member of the Algonquian family and is related to Cree and Naskapi with which it forms a dialectic continuum (Drapeau, 2014) . Statistics Canada estimated the number of speakers at 11,360 in 2016 3 .", | |
| "cite_spans": [ | |
| { | |
| "start": 94, | |
| "end": 108, | |
| "text": "(Mollen, 2006)", | |
| "ref_id": "BIBREF26" | |
| }, | |
| { | |
| "start": 205, | |
| "end": 226, | |
| "text": "(Baraby et al., 2017)", | |
| "ref_id": "BIBREF3" | |
| }, | |
| { | |
| "start": 383, | |
| "end": 398, | |
| "text": "(Drapeau, 2014)", | |
| "ref_id": "BIBREF6" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Innu-Aimun language", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "Although Innu-Aimun is fundamentally an oral language, its orthography was standardized in 1989 (Mollen, 2006) . A first dictionnary based on the standard orthography, for Innu-French, was published in 1991 (Drapeau, 1991) . There exists today a more complete, trilingual and pan-dialectal dictionnary that is being continuously updated and is available online 4 . Other online resources include a verb conjugation web application (Baraby and Junker, 2011) , based on work of Baraby (1998) .", | |
| "cite_spans": [ | |
| { | |
| "start": 96, | |
| "end": 110, | |
| "text": "(Mollen, 2006)", | |
| "ref_id": "BIBREF26" | |
| }, | |
| { | |
| "start": 207, | |
| "end": 222, | |
| "text": "(Drapeau, 1991)", | |
| "ref_id": "BIBREF5" | |
| }, | |
| { | |
| "start": 431, | |
| "end": 456, | |
| "text": "(Baraby and Junker, 2011)", | |
| "ref_id": "BIBREF2" | |
| }, | |
| { | |
| "start": 476, | |
| "end": 489, | |
| "text": "Baraby (1998)", | |
| "ref_id": "BIBREF1" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Innu-Aimun language", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "The aforementioned online tools have been part of an effort by Junker et al. (2016) to develop a series of Web tools for Innu-Aimun language maintenance. This project, primarily, aimed at bilingual speakers, which also includes several primary language resources (e.g. lexicons, grammars, conversational guides, etc.), educational online games and a catalog of audio and written Innu-Aimun works 5 .", | |
| "cite_spans": [ | |
| { | |
| "start": 63, | |
| "end": 83, | |
| "text": "Junker et al. (2016)", | |
| "ref_id": "BIBREF15" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Innu-Aimun language", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "Other than online tools, very few language technologies have been developed for Innu-Aimun, to our knowledge. A search-engine with flexible orthography has been developed by Junker and Stewart (2008) and integrated with an online dictionnary (Junker et al., 2016) , in conjunction with an equivalent tool for East Cree. Other research projects have targeted the construction of Innu-Aimun corpora Cadotte et al. (2022) . Drapeau and Lambert-Br\u00e9ti\u00e8re (2013) proposed an annotated, multimodal corpus with translations. An NRC Canada indigenous languages technology project (Kuhn et al., 2020) aimed to transcribe oral recordings of several indigenous languages in Canada, including Innu-Aimun.", | |
| "cite_spans": [ | |
| { | |
| "start": 174, | |
| "end": 199, | |
| "text": "Junker and Stewart (2008)", | |
| "ref_id": "BIBREF16" | |
| }, | |
| { | |
| "start": 242, | |
| "end": 263, | |
| "text": "(Junker et al., 2016)", | |
| "ref_id": "BIBREF15" | |
| }, | |
| { | |
| "start": 397, | |
| "end": 418, | |
| "text": "Cadotte et al. (2022)", | |
| "ref_id": "BIBREF4" | |
| }, | |
| { | |
| "start": 421, | |
| "end": 456, | |
| "text": "Drapeau and Lambert-Br\u00e9ti\u00e8re (2013)", | |
| "ref_id": "BIBREF7" | |
| }, | |
| { | |
| "start": 571, | |
| "end": 590, | |
| "text": "(Kuhn et al., 2020)", | |
| "ref_id": "BIBREF13" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Innu-Aimun language", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "3 Our proposed approach", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Innu-Aimun language", | |
| "sec_num": "2.2" | |
| }, | |
| { | |
| "text": "In this paper, we focus on the surface segmentation (Ruokolainen et al., 2016; Kann et al., 2018; Liu et al., 2021) , where a term is segmented in a substrings sequence.", | |
| "cite_spans": [ | |
| { | |
| "start": 52, | |
| "end": 78, | |
| "text": "(Ruokolainen et al., 2016;", | |
| "ref_id": "BIBREF29" | |
| }, | |
| { | |
| "start": 79, | |
| "end": 97, | |
| "text": "Kann et al., 2018;", | |
| "ref_id": "BIBREF17" | |
| }, | |
| { | |
| "start": 98, | |
| "end": 115, | |
| "text": "Liu et al., 2021)", | |
| "ref_id": "BIBREF22" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Model overview", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "Given an Innu-Aimun word, the segmentation process consists of breaking down the word into separate morphemes, for example, uminushima \u2192 u-minush-im-a (in English: her/his cats). Our model is made following these steps: (1) apply the Transformer-based encoder-decoder architecture, with a multihead self attention mechanism (Vaswani et al., 2017) ; (2) deal with surface segmentation, while considering the monotonic aspect of morphotactics (that is, the constraints on the ordering of morphemes) (Figure 1) . We train the positional embeddings using the position of each element in a sequence. ", | |
| "cite_spans": [ | |
| { | |
| "start": 324, | |
| "end": 346, | |
| "text": "(Vaswani et al., 2017)", | |
| "ref_id": "BIBREF31" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 497, | |
| "end": 507, | |
| "text": "(Figure 1)", | |
| "ref_id": "FIGREF1" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Model overview", | |
| "sec_num": "3.1" | |
| }, | |
| { | |
| "text": "We model a deep learning-based morphological segmenter using the Transformer-based encoderdecoder architecture. In the encoder, the input sequence is encoded at character level. Then the embedding layer is incorporated with pre-trained embeddings at multiple levels such as character, affix (prefixes and suffixes), along with multiheaded attention over the input sequence, that helps finding morpheme boundaries related to the whole word.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Deep Learning-based morphotactics modeling", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "To ensure the monotonic aspect of morphotactics, the positional embeddings are used to encode the order of each element of a sequence in both the encoder and the decoder.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Deep Learning-based morphotactics modeling", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "The decoder uses the same concept of multihead attention over itself and also the encoder. The attention mechanism allows to align input sequences to the correct corresponding output sequences that are segmented in individual morphemes (Figure 1) .", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 236, | |
| "end": 246, | |
| "text": "(Figure 1)", | |
| "ref_id": "FIGREF1" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Deep Learning-based morphotactics modeling", | |
| "sec_num": "3.2" | |
| }, | |
| { | |
| "text": "A small corpus was manually collected from multiple resources such as the Website of Aimun-Mashinaikan-French-English dictionary Innu 6 as well as open source grammar books and the online Innu lessons platform 7 that are available at the Tshakapesh Institute (Drapeau, 2014; Mollen, 2006) .", | |
| "cite_spans": [ | |
| { | |
| "start": 259, | |
| "end": 274, | |
| "text": "(Drapeau, 2014;", | |
| "ref_id": "BIBREF6" | |
| }, | |
| { | |
| "start": 275, | |
| "end": 288, | |
| "text": "Mollen, 2006)", | |
| "ref_id": "BIBREF26" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Data Preparation", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "The collected experimental corpus contains 500 word bases (roots) and 500 affixes (prefixes, suffixes). A training set, crawled from the Aimun-Mashinaikan dictionary Innu, consists of 30,118 terms, used as raw word lists, non segmented, with length between 2 and 46 characters. A small golden testing set, containing 250 unique terms, was manually segmented with the help of an Innu language teacher from the Uashat Mak Mani-utenam community 8 .", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Data Preparation", | |
| "sec_num": "4.1" | |
| }, | |
| { | |
| "text": "We configured several baselines: (1) based on a simple weighted Finite-State Transducer (FST) to maximise the morpheme frequency (Richardson and Tyers, 2021) , (2) based on Morfessor version 2.0 (Virpioja et al., 2013) to learn the morpheme boundaries using minimum description length optimization, and (3) based on the Adaptor Grammar approach. We used the MorphAGram toolkit (Eskander et al., 2020), with two settings: standard setting (AdaGra-Std) and scholar seeded setting (AdaGra-SS). We adopted the best learning settings: the best standard PrefixStemSuffix+SuffixMorph grammar and the best scholar-seeded grammar, as explained in (Eskander et al., 2019) , for Innu-Aimun.", | |
| "cite_spans": [ | |
| { | |
| "start": 129, | |
| "end": 157, | |
| "text": "(Richardson and Tyers, 2021)", | |
| "ref_id": "BIBREF28" | |
| }, | |
| { | |
| "start": 195, | |
| "end": 218, | |
| "text": "(Virpioja et al., 2013)", | |
| "ref_id": "BIBREF32" | |
| }, | |
| { | |
| "start": 638, | |
| "end": 661, | |
| "text": "(Eskander et al., 2019)", | |
| "ref_id": "BIBREF10" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Training settings", | |
| "sec_num": "4.2" | |
| }, | |
| { | |
| "text": "We configured a deep-learning based model (T-DeepLo) with an encoder-decoder Transformer model (Vaswani et al., 2017) , based entirely on the multihead self-attention mechanism. For the hyperparameters, we used 4-layer both in the encoder and in the decoder. The batch size was set at 32. The initial learning rate was set to 0.0001. The hidden dimension was set at 256, and dropout with a rate of 0.2. The model is trained with 8 multihead attention in the encoder and in the decoder, using Adam optimizer (Kingma and Ba, 2014). The performances of all the models were evaluated using the conventional automatic metrics in the field of NLP, such as Precision, Recall and F1score.", | |
| "cite_spans": [ | |
| { | |
| "start": 95, | |
| "end": 117, | |
| "text": "(Vaswani et al., 2017)", | |
| "ref_id": "BIBREF31" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Training settings", | |
| "sec_num": "4.2" | |
| }, | |
| { | |
| "text": "For the unsupervised methods, we noticed that the scholar-seeded learning (AdaGra-SS) model outperformed all the other baselines, with 70.45%, 61.36%, 65.60% in terms of Precision, Recall and F1 score, respectively (Table 1) . We observed both precision and recall were significantly improved while injecting a list of affixes (prefixes and suffixes) during the training. However, the Morfessor model showed the worst results, with only 40.49% in terms of F1.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 215, | |
| "end": 224, | |
| "text": "(Table 1)", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Precision", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The Transformer-based DeepLo model obtained the best performance across all metrics, with gains of +10.82%, +15.79%, +13.56% in terms of Precision, Recall and F1 score, respectively, compared to the AdaGra-SS model ( Table 1 ). The T-DeepLo model showed the ability to learn and to extract more complex features, relying on the multihead self attention mechanism.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 217, | |
| "end": 224, | |
| "text": "Table 1", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Precision", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "We performed an error analysis in order to shed some light on how the models were able to learn and recognize the morpheme boundary of a sequence. Table 2 shows sample prediction outputs from all the models on the test set.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 147, | |
| "end": 154, | |
| "text": "Table 2", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Precision", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Due to the complex linguistic peculiarities of Innu-Aimun and its dialectal variations, a word can be pronounced in several ways. Thus, its transcription poses more challenges in the segmentation task. Besides, a word in Innu-Aimun is always composed of a central core (root), including a verb.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Error analysis", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "With the help of an Innu language teacher, we made observations and reviewed the data and predictions to determine if the segmentation results were correct and discover the errors. Basically, our models tend to over-segment more complex morphemes due to the linguistic irregularities and the morphotactic phenomena, to detect common lexical suffixes such as ap, tsh or grammatical ending suffixes such as at, eu, t, n, it, mi or uk. In particular, we observed an over-segmentation in the FST and Morfessor models. These models tend to segment a term into several sub-morphemes ( Table 2) . The same phenomena are found in other models of AdaGra-Std and AdaGra-SS. Furthermore, the T-DeepLo model was able to better detect morpheme boundaries.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 579, | |
| "end": 587, | |
| "text": "Table 2)", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Error analysis", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "All models failed when dealing with out-ofvocabulary words. For example, here, the term mitshuap (meaning: house), which was not seen in the training, was segmented into multiple morphemes (Table 2) .", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 189, | |
| "end": 198, | |
| "text": "(Table 2)", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Error analysis", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "Another challenge is related to the oversegmentation of all the models, down to character level, due to the length of prefixes and suffixes between one and multiple characters. For example, some models divided a term up to a character level (Table 2 ): (FST) u a pa tamu; mi t shu a p; (Morfessor) u apa tamu; (AdaGra-Std) minu sha t; (AdaGra-SS) u apa tamu.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 241, | |
| "end": 249, | |
| "text": "(Table 2", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Error analysis", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "We presented a deep learning-based method for morphological segmentation for Innu-Aimun, an indigenous language of Canada, which can be considered as a first research study on the subject, so far.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion and Perspectives", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "Our evaluations showed promising results. Thus, the proposed deep learning-based method, incorporating pre-trained embeddings at multiple levels, helped finding morpheme boundaries related to the whole word. This study makes an important contribution by focusing on morpheme segmentation in the low-resource indigenous language. Furthermore, through this research, we noted the importance of close collaboration and consultation with the Innu indigenous community, to ensure that language technologies are developed with respect and in accordance with the community's revitalisation objectives.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion and Perspectives", | |
| "sec_num": "6" | |
| }, | |
| { | |
| "text": "https://www.thecanadianencyclopedia.ca/en/article/innumontagnais-naskapi 2 https://en.wikipedia.org/wiki/Innu-Aimun", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Statistics Canada: The Aboriginal languages of First Nations people, M\u00e9tis and Inuit 4 https://dictionary.Innu-Aimun.ca/ 5 Tshakapesh Institute -Catalogue", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "https://dictionary.Innu-Aimun.ca/ Words 7 https://lessons.innu.atlas-ling.ca/ 8 https://www.itum.qc.ca/", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [ | |
| { | |
| "text": "The authors are grateful to the community of Uashat Mak Mani-Utenam, Samuel Marticotte for his role in initiating this project, Mrs. Denise Jourdain for sharing her knowledge and experience,", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Acknowledgements", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Morfessor AdaGra-Std AdaGra-SS T-DeepLo akushi\u0144ua akushi\u0144ua akushi\u0144 ua akushi\u0144ua akushi\u0144ua akushi\u0144ua minush at minush at minu sh at minu sha t minush at minush at mi tshuap mi t shu a p mi tsh uap mi tshu ap mi tsh uap mitshuap pashu e u pa shu eu pashueu pa sh ueu pash u eu pashu eu tshika papata n tshi ka pa pa ta n tshi ka papa tan tshika papatan tshika papatan tshika pa patan u minush im a u minush ima u minu sh ima u minush ima u minush im a u minush im a uapat am u u a pa tamu u apa tamu uapa tamu u apa tamu u apa tamu Prof. Yvette Mollen and Prof. Jimena Terraza from Kiuna College, for their precious advices and feedbacks. The authors also thank the anonymous reviewers for their valuable comments.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "FST", | |
| "sec_num": null | |
| } | |
| ], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Converting a comprehensive lexical database into a computational model: The case of East Cree verb inflection", | |
| "authors": [ | |
| { | |
| "first": "Antti", | |
| "middle": [], | |
| "last": "Arppe", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Delasie", | |
| "middle": [], | |
| "last": "Torkornoo", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "Proceedings of the 2nd Workshop on the Use of Computational Methods in the Study of Endangered Languages", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "52--56", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.18653/v1/W17-0108" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Antti Arppe, Marie-Odile Junker, and Delasie Torko- rnoo. 2017. Converting a comprehensive lexical database into a computational model: The case of East Cree verb inflection. In Proceedings of the 2nd Workshop on the Use of Computational Methods in the Study of Endangered Languages, pages 52-56, Honolulu. Association for Computational Linguis- tics.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Guide pratique des principales conjugaisons en Montagnais. Sept-Iles: Institut culturel et \u00e9ducatif montagnais", | |
| "authors": [ | |
| { | |
| "first": "Anne-Marie", | |
| "middle": [], | |
| "last": "Baraby", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Anne-Marie Baraby. 1998. Guide pratique des princi- pales conjugaisons en Montagnais. Sept-Iles: Institut culturel et \u00e9ducatif montagnais.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Conjugaisons des verbes innus", | |
| "authors": [ | |
| { | |
| "first": "Anne-Marie", | |
| "middle": [], | |
| "last": "Baraby", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Anne-Marie Baraby and Marie-Odile Junker. 2011. Conjugaisons des verbes innus.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "A 45-year old language documentation program first aimed at speakers: the case of the Innu", | |
| "authors": [ | |
| { | |
| "first": "Anne-Marie", | |
| "middle": [], | |
| "last": "Baraby", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yvette", | |
| "middle": [], | |
| "last": "Mollen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Anne-Marie Baraby, Marie-Odile Junker, and Yvette Mollen. 2017. A 45-year old language documenta- tion program first aimed at speakers: the case of the Innu.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Challenges and perspectives for innu-aimun within indigenous language technologies", | |
| "authors": [ | |
| { | |
| "first": "Antoine", | |
| "middle": [], | |
| "last": "Cadotte", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Tan", | |
| "middle": [ | |
| "Ngoc" | |
| ], | |
| "last": "Le", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Boivin", | |
| "middle": [], | |
| "last": "Boivin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Fatiha", | |
| "middle": [], | |
| "last": "Sadat", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2022, | |
| "venue": "Proceedings of the Fifth Workshop on the Use of Computational Methods in the Study of Endangered Languages", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "99--108", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Antoine Cadotte, Tan Ngoc Le, Boivin Boivin, and Fatiha Sadat. 2022. Challenges and perspectives for innu-aimun within indigenous language technologies. In Proceedings of the Fifth Workshop on the Use of Computational Methods in the Study of Endangered Languages, pages 99-108, Dublin, Ireland. Associa- tion for Computational Linguistics.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Dictionnaire montagnais-fran\u00e7ais", | |
| "authors": [ | |
| { | |
| "first": "Lynn", | |
| "middle": [], | |
| "last": "Drapeau", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1991, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lynn Drapeau. 1991. Dictionnaire montagnais-fran\u00e7ais. Presses de l'Universit\u00e9 du Qu\u00e9bec.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "Grammaire de la langue innue", | |
| "authors": [ | |
| { | |
| "first": "Lynn", | |
| "middle": [], | |
| "last": "Drapeau", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lynn Drapeau. 2014. Grammaire de la langue innue. Presses de l'Universit\u00e9 du Qu\u00e9bec.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "The innu language documentation project", | |
| "authors": [ | |
| { | |
| "first": "Lynn", | |
| "middle": [], | |
| "last": "Drapeau", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ren\u00e9e", | |
| "middle": [], | |
| "last": "Lambert-Br\u00e9ti\u00e8re", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 17th Foundation for Endangered Languages Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Lynn Drapeau and Ren\u00e9e Lambert-Br\u00e9ti\u00e8re. 2013. The innu language documentation project. In Proceed- ings of the 17th Foundation for Endangered Lan- guages Conference.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "Morphagram, evaluation and framework for unsupervised morphological segmentation", | |
| "authors": [], | |
| "year": null, | |
| "venue": "Proceedings of The 12th Language Resources and Evaluation Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "7112--7122", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Morphagram, evaluation and framework for unsu- pervised morphological segmentation. In Proceed- ings of The 12th Language Resources and Evaluation Conference, pages 7112-7122.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "Unsupervised morphological segmentation for low-resource polysynthetic languages", | |
| "authors": [ | |
| { | |
| "first": "Ramy", | |
| "middle": [], | |
| "last": "Eskander", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Judith", | |
| "middle": [ | |
| "L" | |
| ], | |
| "last": "Klavans", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Smaranda", | |
| "middle": [], | |
| "last": "Muresan", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2019, | |
| "venue": "Proceedings of the 16th Workshop on Computational Research in Phonetics, Phonology, and Morphology", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "189--195", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ramy Eskander, Judith L Klavans, and Smaranda Mure- san. 2019. Unsupervised morphological segmenta- tion for low-resource polysynthetic languages. In Proceedings of the 16th Workshop on Computational Research in Phonetics, Phonology, and Morphology, pages 189-195.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "The uqailaut project", | |
| "authors": [ | |
| { | |
| "first": "Benoit", | |
| "middle": [], | |
| "last": "Farley", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2012, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Benoit Farley. 2012. The uqailaut project. URL http://www.inuktitutcomputing.ca.", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "Learning from the computational modelling of Plains Cree verbs", | |
| "authors": [ | |
| { | |
| "first": "G", | |
| "middle": [], | |
| "last": "Atticus", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Katherine", | |
| "middle": [], | |
| "last": "Harrigan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Antti", | |
| "middle": [], | |
| "last": "Schmirler", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Arppe", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "Morphology", | |
| "volume": "27", | |
| "issue": "4", | |
| "pages": "565--598", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.1007/s11525-017-9315-x" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Atticus G. Harrigan, Katherine Schmirler, Antti Arppe, Lene Antonsen, Trond Trosterud, and Arok Wolven- grey. 2017. Learning from the computational mod- elling of Plains Cree verbs. Morphology, 27(4):565- 598.", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "The Nunavut Hansard Inuktitut-English parallel corpus 3.0 with preliminary machine translation results", | |
| "authors": [ | |
| { | |
| "first": "Eric", | |
| "middle": [], | |
| "last": "Joanis", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Rebecca", | |
| "middle": [], | |
| "last": "Knowles", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Roland", | |
| "middle": [], | |
| "last": "Kuhn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Samuel", | |
| "middle": [], | |
| "last": "Larkin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Patrick", | |
| "middle": [], | |
| "last": "Littell", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Chi-Kiu", | |
| "middle": [], | |
| "last": "Lo", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Darlene", | |
| "middle": [], | |
| "last": "Stewart", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jeffrey", | |
| "middle": [], | |
| "last": "Micher", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2020, | |
| "venue": "Proceedings of the 12th Language Resources and Evaluation Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "2562--2572", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Eric Joanis, Rebecca Knowles, Roland Kuhn, Samuel Larkin, Patrick Littell, Chi-kiu Lo, Darlene Stewart, and Jeffrey Micher. 2020. The Nunavut Hansard Inuktitut-English parallel corpus 3.0 with prelimi- nary machine translation results. In Proceedings of the 12th Language Resources and Evaluation Confer- ence, pages 2562-2572, Marseille, France. European Language Resources Association.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "Adaptor grammars: A framework for specifying compositional nonparametric bayesian models", | |
| "authors": [ | |
| { | |
| "first": "Mark", | |
| "middle": [], | |
| "last": "Johnson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Thomas", | |
| "middle": [], | |
| "last": "Griffiths", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Sharon", | |
| "middle": [], | |
| "last": "Goldwater", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2006, | |
| "venue": "Advances in Neural Information Processing Systems", | |
| "volume": "19", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Mark Johnson, Thomas Griffiths, and Sharon Goldwater. 2006. Adaptor grammars: A framework for speci- fying compositional nonparametric bayesian models. In Advances in Neural Information Processing Sys- tems, volume 19. MIT Press.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "Integrated web tools for Innu language maintenance", | |
| "authors": [ | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yvette", | |
| "middle": [], | |
| "last": "Mollen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H\u00e9l\u00e8ne", | |
| "middle": [], | |
| "last": "St-Onge", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Delasie", | |
| "middle": [], | |
| "last": "Torkornoo", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Papers of the 44th Algonquian Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "192--210", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marie-Odile Junker, Yvette Mollen, H\u00e9l\u00e8ne St-Onge, and Delasie Torkornoo. 2016. Integrated web tools for Innu language maintenance. In Papers of the 44th Algonquian Conference, pages 192-210.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "Building search engines for Algonquian languages. Algonquian Papers-Archive", | |
| "authors": [ | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Terry", | |
| "middle": [], | |
| "last": "Stewart", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "", | |
| "volume": "39", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Marie-Odile Junker and Terry Stewart. 2008. Build- ing search engines for Algonquian languages. Algo- nquian Papers-Archive, 39.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Fortification of neural morphological segmentation models for polysynthetic minimal-resource languages", | |
| "authors": [ | |
| { | |
| "first": "Katharina", | |
| "middle": [], | |
| "last": "Kann", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Manuel", | |
| "middle": [], | |
| "last": "Mager", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ivan", | |
| "middle": [], | |
| "last": "Meza-Ruiz", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Hinrich", | |
| "middle": [], | |
| "last": "Sch\u00fctze", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2018, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1804.06024" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Katharina Kann, Manuel Mager, Ivan Meza-Ruiz, and Hinrich Sch\u00fctze. 2018. Fortification of neural morphological segmentation models for polysyn- thetic minimal-resource languages. arXiv preprint arXiv:1804.06024.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "Adam: A method for stochastic optimization", | |
| "authors": [ | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Diederik", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jimmy", | |
| "middle": [], | |
| "last": "Kingma", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Ba", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2014, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "arXiv": [ | |
| "arXiv:1412.6980" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "blum, and Heather Souter. 2020. The Indigenous Languages Technology project at NRC Canada: An empowerment-oriented approach to developing language software", | |
| "authors": [ | |
| { | |
| "first": "Roland", | |
| "middle": [], | |
| "last": "Kuhn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Fineen", | |
| "middle": [], | |
| "last": "Davis", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Alain", | |
| "middle": [], | |
| "last": "D\u00e9silets", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Eric", | |
| "middle": [], | |
| "last": "Joanis", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Anna", | |
| "middle": [], | |
| "last": "Kazantseva", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Rebecca", | |
| "middle": [], | |
| "last": "Knowles", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Patrick", | |
| "middle": [], | |
| "last": "Littell", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Delaney", | |
| "middle": [], | |
| "last": "Lothian", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Aidan", | |
| "middle": [], | |
| "last": "Pine", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Caroline", | |
| "middle": [ | |
| "Running" | |
| ], | |
| "last": "Wolf", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Eddie", | |
| "middle": [], | |
| "last": "Santos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Darlene", | |
| "middle": [], | |
| "last": "Stewart", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Gilles", | |
| "middle": [], | |
| "last": "Boulianne", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Vishwa", | |
| "middle": [], | |
| "last": "Gupta", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Brian", | |
| "middle": [ | |
| "Maracle" | |
| ], | |
| "last": "Owennat\u00e9kha", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "'", | |
| "middle": [], | |
| "last": "Akwirat\u00e9kha", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Christopher", | |
| "middle": [], | |
| "last": "Martin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Cox", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Olivia", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Sammons", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": null, | |
| "venue": "Proceedings of the 28th International Conference on Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "5866--5878", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.18653/v1/2020.coling-main.516" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Roland Kuhn, Fineen Davis, Alain D\u00e9silets, Eric Joa- nis, Anna Kazantseva, Rebecca Knowles, Patrick Littell, Delaney Lothian, Aidan Pine, Caroline Run- ning Wolf, Eddie Santos, Darlene Stewart, Gilles Boulianne, Vishwa Gupta, Brian Maracle Owen- nat\u00e9kha, Akwirat\u00e9kha' Martin, Christopher Cox, Marie-Odile Junker, Olivia Sammons, Delasie Torko- rnoo, Nathan Thanyeht\u00e9nhas Brinklow, Sara Child, Beno\u00eet Farley, David Huggins-Daines, Daisy Rosen- blum, and Heather Souter. 2020. The Indigenous Languages Technology project at NRC Canada: An empowerment-oriented approach to developing lan- guage software. In Proceedings of the 28th Inter- national Conference on Computational Linguistics, pages 5866-5878, Barcelona, Spain (Online). Inter- national Committee on Computational Linguistics.", | |
| "links": null | |
| }, | |
| "BIBREF20": { | |
| "ref_id": "b20", | |
| "title": "Low-resource NMT: an empirical study on the effect of rich morphological word segmentation on Inuktitut", | |
| "authors": [ | |
| { | |
| "first": "Ngoc", | |
| "middle": [], | |
| "last": "Tan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Fatiha", | |
| "middle": [], | |
| "last": "Le", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Sadat", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2020, | |
| "venue": "Proceedings of the 14th Conference of the Association for Machine Translation in the Americas", | |
| "volume": "1", | |
| "issue": "", | |
| "pages": "165--172", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Tan Ngoc Le and Fatiha Sadat. 2020. Low-resource NMT: an empirical study on the effect of rich mor- phological word segmentation on Inuktitut. In Pro- ceedings of the 14th Conference of the Association for Machine Translation in the Americas (Volume 1: Research Track), pages 165-172, Virtual. Associa- tion for Machine Translation in the Americas (AMTA 2020).", | |
| "links": null | |
| }, | |
| "BIBREF21": { | |
| "ref_id": "b21", | |
| "title": "Indigenous language technologies in canada: Assessment, challenges, and successes", | |
| "authors": [ | |
| { | |
| "first": "Patrick", | |
| "middle": [], | |
| "last": "Littell", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Anna", | |
| "middle": [], | |
| "last": "Kazantseva", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Roland", | |
| "middle": [], | |
| "last": "Kuhn", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Aidan", | |
| "middle": [], | |
| "last": "Pine", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Antti", | |
| "middle": [], | |
| "last": "Arppe", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Christopher", | |
| "middle": [], | |
| "last": "Cox", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Marie-Odile", | |
| "middle": [], | |
| "last": "Junker", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2018, | |
| "venue": "Proceedings of the 27th International Conference on Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "2620--2632", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Patrick Littell, Anna Kazantseva, Roland Kuhn, Aidan Pine, Antti Arppe, Christopher Cox, and Marie-Odile Junker. 2018. Indigenous language technologies in canada: Assessment, challenges, and successes. In Proceedings of the 27th International Conference on Computational Linguistics, pages 2620-2632.", | |
| "links": null | |
| }, | |
| "BIBREF22": { | |
| "ref_id": "b22", | |
| "title": "Morphological segmentation for seneca", | |
| "authors": [ | |
| { | |
| "first": "Zoey", | |
| "middle": [], | |
| "last": "Liu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Robert", | |
| "middle": [], | |
| "last": "Jimerson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Emily", | |
| "middle": [], | |
| "last": "Prud", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "'", | |
| "middle": [], | |
| "last": "Hommeaux", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2021, | |
| "venue": "Proceedings of the First Workshop on Natural Language Processing for Indigenous Languages of the Americas", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "90--101", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Zoey Liu, Robert Jimerson, and Emily Prud'hommeaux. 2021. Morphological segmentation for seneca. In Proceedings of the First Workshop on Natural Lan- guage Processing for Indigenous Languages of the Americas, pages 90-101.", | |
| "links": null | |
| }, | |
| "BIBREF23": { | |
| "ref_id": "b23", | |
| "title": "Probabilistic Finite-State morphological segmenter for Wixarika (huichol) language", | |
| "authors": [ | |
| { | |
| "first": "Manuel", | |
| "middle": [], | |
| "last": "Mager", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Di\u00f3nico", | |
| "middle": [], | |
| "last": "Carrillo", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ivan", | |
| "middle": [], | |
| "last": "Meza", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2018, | |
| "venue": "Journal of Intelligent & Fuzzy Systems", | |
| "volume": "34", | |
| "issue": "5", | |
| "pages": "3081--3087", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.3233/JIFS-169492" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Manuel Mager, Di\u00f3nico Carrillo, and Ivan Meza. 2018a. Probabilistic Finite-State morphological segmenter for Wixarika (huichol) language. Journal of Intelli- gent & Fuzzy Systems, 34(5):3081-3087. Publisher: IOS Press.", | |
| "links": null | |
| }, | |
| "BIBREF24": { | |
| "ref_id": "b24", | |
| "title": "Challenges of language technologies for the indigenous languages of the Americas", | |
| "authors": [ | |
| { | |
| "first": "Manuel", | |
| "middle": [], | |
| "last": "Mager", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ximena", | |
| "middle": [], | |
| "last": "Gutierrez-Vasques", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Gerardo", | |
| "middle": [], | |
| "last": "Sierra", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Ivan", | |
| "middle": [], | |
| "last": "Meza-Ruiz", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2018, | |
| "venue": "Proceedings of the 27th International Conference on Computational Linguistics", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "55--69", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Manuel Mager, Ximena Gutierrez-Vasques, Gerardo Sierra, and Ivan Meza-Ruiz. 2018b. Challenges of language technologies for the indigenous languages of the Americas. In Proceedings of the 27th Inter- national Conference on Computational Linguistics, pages 55-69, Santa Fe, New Mexico, USA. Associa- tion for Computational Linguistics.", | |
| "links": null | |
| }, | |
| "BIBREF25": { | |
| "ref_id": "b25", | |
| "title": "Bootstrapping a neural morphological generator from morphological analyzer output for inuktitut", | |
| "authors": [ | |
| { | |
| "first": "Jeffrey", | |
| "middle": [], | |
| "last": "Micher", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2019, | |
| "venue": "Proceedings of the Workshop on Computational Methods for Endangered Languages", | |
| "volume": "2", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jeffrey Micher. 2019. Bootstrapping a neural morpho- logical generator from morphological analyzer out- put for inuktitut. In Proceedings of the Workshop on Computational Methods for Endangered Languages, volume 2, page 7.", | |
| "links": null | |
| }, | |
| "BIBREF26": { | |
| "ref_id": "b26", | |
| "title": "Transmettre un h\u00e9ritage: la langue innue. Cap-aux-Diamants: la revue d'histoire du Qu\u00e9bec", | |
| "authors": [ | |
| { | |
| "first": "Yvette", | |
| "middle": [], | |
| "last": "Mollen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2006, | |
| "venue": "Publisher: Les \u00c9ditions Capaux-Diamants inc", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "21--25", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yvette Mollen. 2006. Transmettre un h\u00e9ritage: la langue innue. Cap-aux-Diamants: la revue d'histoire du Qu\u00e9bec, (85):21-25. Publisher: Les \u00c9ditions Cap- aux-Diamants inc.", | |
| "links": null | |
| }, | |
| "BIBREF27": { | |
| "ref_id": "b27", | |
| "title": "Morphological word segmentation on agglutinative languages for neural machine translation", | |
| "authors": [ | |
| { | |
| "first": "Yirong", | |
| "middle": [], | |
| "last": "Pan", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Xiao", | |
| "middle": [], | |
| "last": "Li", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Yating", | |
| "middle": [], | |
| "last": "Yang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Rui", | |
| "middle": [], | |
| "last": "Dong", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2020, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.48550/ARXIV.2001.01589" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Yirong Pan, Xiao Li, Yating Yang, and Rui Dong. 2020. Morphological word segmentation on agglutinative languages for neural machine translation.", | |
| "links": null | |
| }, | |
| "BIBREF28": { | |
| "ref_id": "b28", | |
| "title": "A morphological analyser for k'iche", | |
| "authors": [ | |
| { | |
| "first": "Ivy", | |
| "middle": [], | |
| "last": "Richardson", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Francis", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Tyers", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2021, | |
| "venue": "Procesamiento del Lenguaje Natural", | |
| "volume": "66", | |
| "issue": "", | |
| "pages": "99--109", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ivy Richardson and Francis M Tyers. 2021. A mor- phological analyser for k'iche'. Procesamiento del Lenguaje Natural, 66:99-109.", | |
| "links": null | |
| }, | |
| "BIBREF29": { | |
| "ref_id": "b29", | |
| "title": "A comparative study of minimally supervised morphological segmentation", | |
| "authors": [ | |
| { | |
| "first": "Oskar", | |
| "middle": [], | |
| "last": "Teemu Ruokolainen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Kairit", | |
| "middle": [], | |
| "last": "Kohonen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Stig-Arne", | |
| "middle": [], | |
| "last": "Sirts", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Mikko", | |
| "middle": [], | |
| "last": "Gr\u00f6nroos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Sami", | |
| "middle": [], | |
| "last": "Kurimo", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Virpioja", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Computational Linguistics", | |
| "volume": "42", | |
| "issue": "1", | |
| "pages": "91--120", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Teemu Ruokolainen, Oskar Kohonen, Kairit Sirts, Stig- Arne Gr\u00f6nroos, Mikko Kurimo, and Sami Virpioja. 2016. A comparative study of minimally supervised morphological segmentation. Computational Lin- guistics, 42(1):91-120.", | |
| "links": null | |
| }, | |
| "BIBREF30": { | |
| "ref_id": "b30", | |
| "title": "Neural machine translation of rare words with subword units", | |
| "authors": [ | |
| { | |
| "first": "Rico", | |
| "middle": [], | |
| "last": "Sennrich", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Barry", | |
| "middle": [], | |
| "last": "Haddow", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Alexandra", | |
| "middle": [], | |
| "last": "Birch", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2016, | |
| "venue": "Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics", | |
| "volume": "1", | |
| "issue": "", | |
| "pages": "1715--1725", | |
| "other_ids": { | |
| "DOI": [ | |
| "10.18653/v1/P16-1162" | |
| ] | |
| }, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. Neural machine translation of rare words with subword units. In Proceedings of the 54th Annual Meeting of the Association for Computational Lin- guistics (Volume 1: Long Papers), pages 1715-1725, Berlin, Germany. Association for Computational Lin- guistics.", | |
| "links": null | |
| }, | |
| "BIBREF31": { | |
| "ref_id": "b31", | |
| "title": "Attention is all you need", | |
| "authors": [ | |
| { | |
| "first": "Ashish", | |
| "middle": [], | |
| "last": "Vaswani", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Noam", | |
| "middle": [], | |
| "last": "Shazeer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Niki", | |
| "middle": [], | |
| "last": "Parmar", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Jakob", | |
| "middle": [], | |
| "last": "Uszkoreit", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Llion", | |
| "middle": [], | |
| "last": "Jones", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Aidan", | |
| "middle": [ | |
| "N" | |
| ], | |
| "last": "Gomez", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "\u0141ukasz", | |
| "middle": [], | |
| "last": "Kaiser", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Illia", | |
| "middle": [], | |
| "last": "Polosukhin", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2017, | |
| "venue": "Advances in neural information processing systems", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "5998--6008", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, \u0141ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in neural information pro- cessing systems, pages 5998-6008.", | |
| "links": null | |
| }, | |
| "BIBREF32": { | |
| "ref_id": "b32", | |
| "title": "Morfessor 2.0: Python implementation and extensions for morfessor baseline", | |
| "authors": [ | |
| { | |
| "first": "Sami", | |
| "middle": [], | |
| "last": "Virpioja", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Peter", | |
| "middle": [], | |
| "last": "Smit", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Stig-Arne", | |
| "middle": [], | |
| "last": "Gr\u00f6nroos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Mikko", | |
| "middle": [], | |
| "last": "Kurimo", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Sami Virpioja, Peter Smit, Stig-Arne Gr\u00f6nroos, Mikko Kurimo, et al. 2013. Morfessor 2.0: Python imple- mentation and extensions for morfessor baseline.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure", | |
| "text": "Finite-State Transducers (FST), have been used the most.Farley (2012) proposed an FST-based morphological analyser for Inuktitut (one of Canada's most resourced and documented indigenous languages). Harrigan et al. (2017) developed an FST morphological model for Plains Cree. applied the same approach partially adapted to East Cree.Mager et al. (2018a) proposed a probabilistic approach to an FST model for Wixarika (huichol)." | |
| }, | |
| "FIGREF1": { | |
| "num": null, | |
| "uris": null, | |
| "type_str": "figure", | |
| "text": "Architecture of our framework: Deep Learning-based Morphological segmentation for indigenous language, with pretrained debiased word-based embedding for source-target, and positional embedding." | |
| } | |
| } | |
| } | |
| } |