Search

CLARIN is a pan-European Research Infrastructure, which provides accessibility to all European digital language resources and tools through a single sign-on online environment. The CLARIN community offers long-term solutions and technology services for deploying, connecting, analyzing and sustaining digital language data (e.g. various types of corpora, dictionaries, glossaries) and tools (e.g. corpus management and exploration systems, speech processing systems, NLP pipelines). CLARIN supports scholars who want to engage in pioneering data-driven research, contributing to a multilingual European Research Area. In addition, CLARIN is also an infrastructural initiative that is fully committed to the European agenda towards Open Science. By offering resources in open access and promoting the curation and depositing of data in alignment with the requirements for the interoperability of data and services, CLARIN has paved the way for large-scale data sharing and increased reuse of resources. In Switzerland, the CLARIN-CH Consortium represents an inter-university national network of scholars working on language related issues in various sub-disciplines, including Dialectology, Corpus Linguistics, Applied Linguistics, Natural Language Processing, Machine Translation, Speech Technology, etc. CLARIN-CH aims to bring together the research community using language resources, foster the sharing of resources and expertise, encourage collaborations by implementing focused inter-university Working Groups, and give Swiss researchers access to CLARIN’s resources, tools, services and research network. By getting actively involved in CLARIN-CH and CLARIN Europe, Swiss scholars will benefit of: (i) an increased visibility of their assets (research projects, data, tools, methodologies, expertise, etc.), (ii) a better access to European digital language resources and cutting-edge language technologies, (iii) involvement in European collaborative research projects, (iv) access to numerous funding opportunities (workshops, summer schools, mobility grants, etc.).

This report on Europe's Sign Languages is part of a series of language deliverables developed within the framework of the European Language Equality (ELE) project. The series seeks to not only delineate the current state of affairs for each European language, but to additionally identify the gaps and factors that hinder further development in research and technology. The survey presented here focuses on the condition of Language Technology (LT) with regard to Europe's Sign Languages, a set of languages often forgotten in the context of European Language Equality. With the rise of the deep learning paradigm in artificial intelligence, sign language technologies become technologically feasible, provided that enough data is available to feed this data-hungry paradigm. It is exactly the quality and quantity of data that is the main bottleneck in development of well performing and useful technologies. In the past, there have been several projects aimed at developing sign language technologies and methodologies that have been deemed of little value by the deaf communities. Co-creation and involvement of deaf communities throughout projects and development of technologies ensures that this does not happen again., This report was a supplementary contribution to the European Language Equality (ELE) project. It involved authors not associated with ELE and is not a contractual project deliverable.

Poster at SwissText 2023, Neuchâtel, June 2023 In this poster, we showcase how CLARIN – a well-established research infrastructure (RI) and a landmark of the European Strategy Forum on RIs ESFRI – and its newly established Swiss node CLARIN-CH, support the scientific community using language data in the context of Open Science. CLARIN is a pan-European Research Infrastructure (RI), which provides accessibility to all European digital language resources and tools through a single sign-on online environment. CLARIN centres share with the community their expertise, their language resources, tools, and provide a large array of services. Specifically, the CLARIN community offers long-term solutions and technology services for deploying, connecting, analysing and sustaining digital language data and tools. To connect a variety of resources and data that are available, certain metadata norms were developed – component metadata (CMDI). CMDI revolves around XML schemas to make resources easily searchable and adaptable for automated metadata parsing. The XML schemas are stringent enough to make them usable for general purposes but also flexible enough to adapt to each case. Furthermore, CLARIN supports scholars who want to engage in pioneering data-driven research, contributing to a multilingual European Research Area. An example is the ParlaMint corpus, which contributes to the creation of comparable and uniformly annotated multilingual corpora (more than 17 languages) of parliamentary sessions. Finally, by offering resources in OA and promoting the curation and depositing of data in alignment with the requirements for the interoperability of data and services, CLARIN paved the way for large-scale data sharing and increased reuse of resources. In Switzerland, CLARIN-CH is a cross-institutional national network of scholars working on language related issues in various sub-disciplines, which started in December 2020. Its governance is ensured by the CLARIN-CH consortium of Swiss higher education and research institutions, which is fully funded by the member institutions. The consortium works in partnership with the Linguistic Research Infrastructure LiRI (UZH), which provides data and language technology services, and the Language Repository of Switzerland (LaRS/SWISSUbase), which is a national FAIR-complaint repository for linguistic data. Among the objectives of the CLARIN-CH Consortium are: (i) to bring together the research community using language resources and encourage collaborations by implementing cross-institutional working groups (WGs), such as Open Research Data projects for linguistics data and Management of sensitive data and legal aspects for linguistic data in Switzerland, and to foster the federation of resources and expertise at the national and the European levels. Regarding the former, CLARIN-CH WGs target specific needs of individual sub-areas with respect to the development of digital language resources and analysis tools, and their incorporation into the CLARIN-CH infrastructure. For example, the UpLORD project (leading house: UZH) focuses on: (I) upgrading workflows and interoperability of Swiss RIs for linguistic data, i.e., LiRI and SWISSUbase/LaRS, (ii) documenting and promoting best practices, (iii) raising awareness and training about ORD practices in the context of teaching, research and publishing, and (iv) building a robust practice of data curation. Regarding the latter objective, in collaboration with LiRI, CLARIN-CH works towards the interoperability of Swiss corpora with the European CLARIN infrastructure by developing a metadata schema compatible with the CMDI, the construction of an OA national corpus platform to allow complex queries in a variety of corpora and of an annotation tool for multimodal data, and targets the establishment of knowledge centers to value Swiss scholars’ expertise in language technology and language sciences. CLARIN and its Swiss node CLARIN-CH collaborate with other national and European RIs, such as DARIAH and its Swiss node DARIAH-CH, to further develop the Social Sciences and Humanities cluster and represent it in the European Open Science Cloud EOSC. By getting actively involved in CLARIN, Swiss scholars benefit of: (i) an increased visibility of their assets (research projects, data, tools, methodologies, expertise, etc.), (ii) a better access to European digital language resources and cutting-edge language technologies, (iii) involvement in European and national collaborative research projects, (iv) access to numerous funding opportunities (call for EU-funded projects, workshops, summer schools, mobility grants).

For machine translation (MT) systems to produce accurate and fluent translations, reliable parallel corpora are key. Errors, due to misalignments or inadequate filtering during compilation of a parallel corpus, can have detrimental effects on the performance of an MT system trained on the data. Moreover, when the corpus is too small, the MT system may not be able to capture the complexities of the source and target languages and produce accurate translations. However, obtaining high-quality parallel data is often a challenging task, even more so for languages with a low number of speakers or rich morphology exacerbating the data sparsity problem. It is thus imperative to develop accurate methods for processing parallel corpora that can help make the most of what is available. In this thesis, we address this challenge by exploring various methods for processing parallel corpora to maximize their usefulness for MT. First, we investigate a variety of classifiers and scoring mechanisms used for filtering parallel corpora, looking into how efficient they are at removing data detrimental to MT training and retaining useful data. We find that different filtering strategies suit different datasets and that filtering separately for different translation directions can yield better translations in downstream MT tasks. Second, we examine different approaches to sentence alignment, compare their effectiveness, and show that combining multiple methods can improve alignment accuracy. Third, we experiment with comparable corpora mining methods to extract even more useful data from sentences that had previously been discarded, showing that this often overlooked data is a potential source of useful training data. Finally, we manually evaluate translations generated by MT systems trained on our processed datasets, most suitable for each translation direction, confirming the advantages of our applied methods. Our findings highlight the importance of careful processing and curation of parallel corpora for MT. We propose approaches for maximizing the utility of available parallel data, particularly for scenarios where resources are scarce, contributing to the development of more accurate and reliable MT systems., Áreiðanlegar samhliða málheildir eru lykillinn að því að hægt sé að þjálfa þýðingarvélar, sem geta myndað nákvæmar þýðingar sem flæða vel á markmálinu. Skekkjur í þjálfunargögnum, sem koma til vegna rangrar samröðunar setninga eða ófullnægjandi síunar við smíði samhliða málheilda, geta spillt gæðum þýðingarvélar sem þjálfuð er á gögnunum. Of lítil samhliða málheild getur jafnframt orðið til þess að þýðingarvélin nái ekki tökum á málfræði eða öðrum blæbrigðum frum- og markmálanna og myndi þess vegna ónákvæmar þýðingar.Það getur hins vegar verið flókið og erfitt að tryggja hámarksgæði þjálfunargagna við úrvinnslu samhliða texta, ekki síst þegar um er að ræða texta á tungumálum sem fáir tala eða þegar flóknar beygingar og virk orðmyndun auka á vandann við að greina rýr gögn. Þegar samhliða málheildir eru settar saman er því afar mikilvægt að þróa nákvæmar aðferðir sem miða að því að nýta sem allra best þau gögn sem til eru Í þessari ritgerð tökumst við á við þetta vandamál með því að kanna ýmsar aðferðir til að vinna gögn við smíði samhliða málheilda með það að leiðarljósi að hámarka notagildi gagnanna fyrir vélþýðingar. Í fyrsta lagi rannsökum við nokkrar gerðir flokkara og matsaðferðir sem notaðar eru til að sía samhliða málheildir. Við skoðum hversu árangursríkar þær eru til að fjarlægja setningapör sem geta dregið úr gæðum þýðingarvéla ef þau eru hluti þjálfunargagna og hversu líklegar aðferðirnar eru til að halda eftir þeim setningapörum sem búast má við að séu best fallnar til að bæta þýðingarvélarnar. Við komumst að því að mismunandi síunaraðferðir henta mismunandi gagnasöfnun og að með því að sía sérstaklega fyrir hverja þýðingarátt má bæta gæði þýðinga þeirra véla sem þjálfaðar eru á gögnunum. Í öðru lagi skoðum við mismunandi aðferðir við samröðun setninga, berum saman markvirkni þeirra og sýnum að með því að láta margar mismunandi aðferðir vinna saman getum við aukið nákvæmni samröðunarinnar. Í þriðja lagi gerum við tilraunir með aðferðir til að vinna samhliða gögn úr sambærilegum málheildum, og beitum þeim aðferðum til að draga nýtileg gögn úr setningum og setningapörum sem hafnað hefur verið á fyrri stigum í smíði þjálfunargagnanna. Við sýnum með nokkrum tilraunum að mögulegt er að nýta þessi gögn, sem yfirleitt er litið fram hjá, til að stækka samhliða þjálfunarmálheildir með nýtilegum gögnum og þar með bæta þýðingarvélar sem þjálfaðar eru á þeim. Að lokum metum við handvirkt þýðingar myndaðar af þýðingarvélum sem þjálfaðar eru á gögnum sem unnin hafa verið með okkar aðferðum, en það mat staðfestir gagnsemi þeirra aðferða sem við beitum. Niðurstöður okkar undirstrika mikilvægi vandaðrar greiningar og gagnavinnslu við smíði samhliða málheilda sem notaðar eru til að þjálfa þýðingarvélar. Við kynnum aðferðir sem hámarka notagildi tiltækra samhliða gagna, ekki sístþegar takmarkað magn gagna er fyrir hendi, og stuðlum þannig að þróun nákvæmari og áreiðanlegri þýðingarvéla.

Lan honetan, Elhuyar hiztegien lanketan eskuratu dugun esperientziatik abiatuta, euskarazko hiztegigintzak izan duen ibilbidearen zertzelada batzuk emango ditugu, eta aro digitalak eta Interneten aroak hiztegigintza-lanean eta produktu lexikografikoetan ekarri dituzten aldaketak izango ditugu hizketagai. Lehen zatian, euskarazko hiztegigintzaren bilakaera azalduko dugu. Ondoren, teknologia berriek hiztegigintzan ekarri duten eraldaketa izango dugu hizpide; hiztegiak lantzeko metodologia eta tresnetan sakonduko dugu, eta Internetek hiztegien edukia erakusteko moduan, eguneratze-lanetan eta erabiltzaileen beharretan eragin dituen aldaketak aipatuko ditugu. Azkenik, aro digital berri honek dituen alde ilun batzuei buruz arituko gara; izan ere, hiztegigintza-arloko erakunde eta eragile askoren kezka da euskarak ezinbesteko dituen proiektu hauen biziraupena ziurtatzeko baliabide ekonomikoen falta. HITZ GAKOAK: hiztegigintza, digitalizazioa, hizkuntza-teknologiak

Artikulu honetan morfologia konputazionalaren garapena azalduko dugu, Ixa taldeak euskararako egindako aplikazioa azpimarratuz. Bilakaera historikoa jaso nahi izan dugu, teknologiaren bilakaera eta aplikazioen bilakaera uztartuz, beti euskararen gainean egindakoa adibidetzat hartuta. Teknologiaren bilakaerari dagokionez, ezagutzan oinarritutako ereduak ohikoak eta nagusi izan dira morfologian, baina hurbilpen estatistikoan, eta, batez ere, ikasketa sakonean oinarritutakoak, ugalduz joan dira azken urteetan, horietako asko datuak eta gramatikak eskuragarriak ez dituzten hizkuntzetarako (edo dialektoetarako). Aplikazioak direla eta, ezagunena eta erabiliena zuzentzaile ortografikoa da, baina irakaskuntzarekin edo bertsogintzarekin lotutakoak ere aipagarriak dira. HITZ GAKOAK: morfologia konputazionala, zuzenketa ortografikoa, hizkuntz teknologia

In language technology and language sciences, tab-separated values (TSV) represent a frequently used formalism to represent linguistically annotated natural language, often addressed as "CoNLL formats". A large number of such formats do exist, but although they share a number of common features, they are not interoperable, as different pieces of information are encoded differently in these dialects. CoNLL-RDF refers to a programming library and the associated data model that has been introduced to facilitate processing and transforming such TSV formats in a serialization-independent way. CoNLL-RDF represents CoNLL data, by means of RDF graphs and SPARQL update operations, but so far, without machine-readable semantics, with annotation properties created dynamically on the basis of a user-defined mapping from columns to labels. Current applications of CoNLL-RDF include linking between corpora and dictionaries [Mambrini and Passarotti, 2019] and knowledge graphs [Tamper et al., 2018], syntactic parsing of historical languages [Chiarcos et al., 2018; Chiarcos et al., 2018], the consolidation of syntactic and semantic annotations [Chiarcos and Fäth, 2019], a bridge between RDF corpora and a traditional corpus query language [Ionov et al., 2020], and language contact studies [Chiarcos et al., 2018]. We describe a novel extension of CoNLL-RDF, introducing a formal data model, formalized as an ontology. The ontology is a basis for linking RDF corpora with other Semantic Web resources, but more importantly, its application for transformation between different TSV formats is a major step for providing interoperability between CoNLL formats., OASIcs, Vol. 93, 3rd Conference on Language, Data and Knowledge (LDK 2021), pages 20:1-20:14

In an era of rapid digital convergence, a wide web of digital environments, platforms, complex technologies, and services – such as blogs, social networking applications, Wikis, learning management systems, online games, 3D role-playing games, and simulations – overwhelms everyday life. us. Several of these new digital technologies affect both young user-teachers' learning maturation and older teacher-users' training. Today, digital technologies are part of the educational process and dynamically expand the possibilities of both the teacher in providing knowledge and the learner in acquiring knowledge. This volume seeks to introduce students/readers to the radical upheavals that digital technologies suggest for literacy, the cultural practices of reading and writing, and, in particular, for learning a foreign language/culture. In the context of contemporary digital humanities studies, which are currently at the center of the debate to strengthen their presence in the academic and social space, the handbook on digital technologies in foreign language/culture teaching has a theoretical as well as practical orientation. At the level of theory, it clarifies and analyzes basic concepts of modern digital technologies, explains their theoretical origins, and presents their advantages and disadvantages in language teaching issues. On a practical level, it presents a wide range of digital tools/applications available to the teacher of a foreign, second, or native language, to develop modern language activities of comprehension and production of spoken and written language in his students. All suggested digital tools and their applications, grouped appropriately, are accompanied by specific usage scenarios describing their utilization in detail., Σε μια εποχή γρήγορης ψηφιακής σύγκλισης, ένα ευρύ πλέγμα ψηφιακών περιβαλλόντων, πλατφορμών, σύνθετων τεχνολογιών και υπηρεσιών –όπως τα ιστολόγια, οι εφαρμογές κοινωνικής δικτύωσης, τα Wikis, τα συστήματα διαχείρισης μάθησης, τα διαδικτυακά παιχνίδια τα τρισδιάστατα παιχνίδια ρόλων και προσομοιώσεων – κατακλύζει την καθημερινότητά μας. Αρκετές από αυτές τις νέες ψηφιακές τεχνολογίες επηρεάζουν τόσο τη μαθησιακή ενηλικίωση των νεαρών χρηστών-διδασκομένων όσο και την κατάρτιση των μεγαλύτερων-διδασκόντων. Σήμερα οι ψηφιακές τεχνολογίες εντάσσονται στην εκπαιδευτική διαδικασία και επεκτείνουν δυναμικά τις δυνατότητες τόσο του διδάσκοντα στην παροχή γνώσεων, όσο και του διδασκόμενου στην κατάκτηση της γνώσης. Το παρόν σύγγραμμα επιδιώκει να εισαγάγει τους φοιτητές / αναγνώστες στις ριζικές ανατροπές που οι ψηφιακές τεχνολογίες εισηγούνται για την εγγραμματοσύνη, τις πολιτισμικές πρακτικές της ανάγνωσης και της γραφής και, ειδικότερα, για την εκμάθηση μιας ξένης γλώσσας/πολιτισμού. Στο πλαίσιο των σύγχρονων ψηφιακών ανθρωπιστικών σπουδών, που σήμερα βρίσκονται στο επίκεντρο της συζήτησης για την ενίσχυση της παρουσίας τους στον ακαδημαϊκό και κοινωνικό χώρο, το εγχειρίδιο για τις ψηφιακές τεχνολογίες στην διδασκαλία της ξένης γλώσσας/πολιτισμού έχει έναν θεωρητικό όσο και πρακτικό προσανατολισμό. Σε επίπεδο θεωρίας, αποσαφηνίζει και αναλύει βασικές έννοιες των σύγχρονων ψηφιακών τεχνολογιών, εξηγεί τις θεωρητικές καταβολές τους και παρουσιάζει τα πλεονεκτήματα και μειονεκτήματά τους σε ζητήματα γλωσσικής διδασκαλίας. Σε πρακτικό επίπεδο, παρουσιάζει ένα ευρύ φάσμα ψηφιακών εργαλείων/εφαρμογών που έχει στην διάθεσή του ο διδάσκων της ξένης, δεύτερης ή μητρικής γλώσσας, για να αναπτύξει σύγχρονες γλωσσικές δραστηριότητες κατανόησης και παραγωγής προφορικού και γραπτού λόγου στους μαθητές του. Όλα τα προτεινόμενα ψηφιακά εργαλεία και οι εφαρμογές τους, ομαδοποιημένα κατάλληλα, συνοδεύονται από συγκεκριμένα σενάρια χρήσης που περιγράφουν αναλυτικά την αξιοποίησή τους.

The EuLTDom2023 project reports on the current state-of-the-art of the usage of LT in different domains. The main purpose of this deliverable is to map the usage of NLP in various domains, to report findings regarding the fields that make regular use of these technologies, and to list domains that infrequently use LT or do not use it at all. For this aim, we analysed scientific papers published between 2010 and 2022 in the ACL Anthology, thus corresponding to scientific work done by the LT community. By applying a dictionary-based approach based on precise lists of terms related to languages, domains, and NLP tasks, we were able to present an overview of each of these dimensions, and to provide language- specific results mapping the usage of NLP tasks by the different domains. With the overall analysis, it is possible to identify the language, domains, and LT that have been the focus of the NLP research in the past years, and the language-specific results allow the clear identification of potential future developments to improve language equality at the level usage of LT.

High-performance computing is an essential service required in multiple research areas. This report performs desk-research and an online survey to map the HPC landscape avail- able to LT researchers. The report enumerates the various infrastructures available as well as presents the views of the LT researchers obtained via an online survey. In Section 2, background knowledge related to HPC is presented. This information is segmented into six sections and describes HPC along with associated terminologies, HPC classification schemes, HPC initiatives, and aspects related to HPC access and users. In Section 3, methodology re- lated to desk research and an online survey is described. In Section 4, the HPC landscape is described using the information obtained from desk research. In Section 7, responses from the online survey are analysed. The report concludes with recommendations, conclusion, and limitations.

Preprint, The rapid advancements in large language models (LLM) and artificial intelligence (AI) have been a subject of recent significant interest and debate. This paper explores the impact of these developments on language learning. I discuss the technology underlying AI-based tools and the natural language processing (NLP) tasks they were originally designed for. This will help us to identify opportunities and limitations regarding their use in the context of language learning. I then examine how such technology can be used efficiently and effectively in language teaching and learning. The availability of such tools will require language teaching to focus on the non-mechanical aspects of writing. Similarly, automatically produced personalized teaching and learning materials will not replace human teachers, but give space for and support human–human interaction.

The question whether some second language (L2) features (forms, items, structures, patterns, rules) and, if so, which types of features, are more affected by instruction or are more effectively instructed than others has received increasing attention in applied linguistics and SLA research since the mid-1990s (VanPatten, 1994; Spada, 1997). For SLA researchers, knowing how the nature of the input stimulus interacts with the human cognitive processing and learning mechanisms is important for constructing a comprehensive theory of language acquisition. For theoretical, typological and psycho-linguists, knowing which types of features of language are hard to learn and teach can inform theories of language and processing complexity (Szmrecsanyi & Kortmann, 2012). For L2 teachers and curriculum and textbook developers, knowing which type of L2 features to select in the typically limited teaching time is important for optimizing L2 teaching and learning effectiveness (Doughty & Williams, 1998).

As a relatively young field of research and development that began with work on crypt-analysis and machine translation around 50 years ago, natural language processing (NLP) is concerned with the automated processing of human language. Keywords: CALL; computational linguistics; corpora; ICALL; natural language processing