Hasil untuk "Computational linguistics. Natural language processing"

Deyu An, Nan Wang

Abstract In the era of medical big data, the analysis of orthopedic disease types demands precision and multi-dimensional insight. Traditional analysis models, limited by their one-dimensional or simplistic structures, struggle to interpret complex clinical data comprehensively. To address this, we innovatively integrate the entropy weight technique with the Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS) method and the Rank Sum Ratio (RSR) method. This approach leverages objective weighting based on information entropy and precise solution differentiation from the TOPSIS method to enhance the RSR method's ranking capability while retaining the original quantitative information. Based on a dataset of 100 orthopedic diseases across 18 indicators from five evaluation dimensions, the proposed model identified “lumbar disc herniation with nerve root disease” as the top-ranking disease (comprehensive index, Ci = 0.6915), followed by “lumbar disc herniation with sciatica” (Ci = 0.6860) and “lumbar spinal stenosis” (Ci = 0.6368). The functional positioning dimension contributed the highest indicator weight (58.107%), with the “number of patients undergoing level-four surgeries” being the most influential indicator (21.178%). The RSR regression model demonstrated strong goodness of fit (coefficient of determination, R2 = 0.887; F = 770.543; p < 0.001), confirming the robustness and validity of the analysis. This integrated model provides a reliable and objective decision-making tool for classifying orthopaedic disease types in clinical practice. Graphical abstract

Дана Радван

У статті проаналізовано записи хронік про використання меча у середньовічній Русі в метафоричному контексті. Факти, наведені у Повісті врем’яних літ і Радзивілівському літописі, бууть розглянуті у зв’язку із записом Ламперта з Херсфельду про дар угорської королеви Анастасії герцогу Баварії Отто у другій половині 11 ст., меч Аттіли, що дозволить інтерпретувати  метафоричні повідомлення, які передавалися за допомогою меча, і пояснити незвичний церемоніальний вибір угорської королеви київського походження.

Changye Li, Jacob Solinsky, Trevor Cohen et al.

Introduction: While linguistic retrogenesis has been extensively investigated in the neuroscientific and behavioral literature, there has been little work on retrogenesis using computerized approaches to language analysis. Methods: We bridge this gap by introducing a method based on comparing output of a pre-trained neural language model (NLM) with an artificially degraded version of itself to examine the transcripts of speech produced by seniors with and without dementia and healthy children during spontaneous language tasks. We compare a range of linguistic characteristics including language model perplexity, syntactic complexity, lexical frequency and part-of-speech use across these groups. Results: Our results indicate that healthy seniors and children older than 8 years share similar linguistic characteristics, as do dementia patients and children who are younger than 8 years. Discussion: Our study aligns with the growing evidence that language deterioration in dementia mirrors language acquisition in development using computational linguistic methods based on NLMs. This insight underscores the importance of further research to refine its application in guiding developmentally appropriate patient care, particularly in early stages.

Anton Eklund, Mona Forsman, Frank Drewes

Document clustering is frequently used in applications of natural language processing, e.g. to classify news articles or creating topic models. In this paper, we study document clustering with the common clustering pipeline that includes vectorization with BERT or Doc2Vec, dimension reduction with PCA or UMAP, and clustering with K-Means or HDBSCAN. We discuss the inter- actions of the different components in the pipeline, parameter settings, and how to determine an appropriate number of dimensions. The results suggest that BERT embeddings combined with UMAP dimension reduction to no less than 15 dimensions provides a good basis for clustering, regardless of the specific clustering algorithm used. Moreover, while UMAP performed better than PCA in our experiments, tuning the UMAP settings showed little impact on the overall performance. Hence, we recommend configuring UMAP so as to optimize its time efficiency. According to our topic model evaluation, the combination of BERT and UMAP, also used in BERTopic, performs best. A topic model based on this pipeline typically benefits from a large number of clusters.

MUKENGE KABEYA Dédé, MULENDA SUMBU Ali & NTUMBA KONSOKUBI Pierre

Résumé : L’objectif de cet article est d’étudier la prise en charge scolaire des enfants en situation de handicap mental modéré et sévère de Village Bondeko (école spéciale) à travers l’approche orthopédagogique du Programme Educatif Individualisé (PEI). C’est dans ce contexte que nous avons analysé 6 cas des dossiers de Programme Educatif Individualisé (PEI) des élèves en situation de handicap mental modéré et sévère évalués par l’équipe éducative de cette école. En partant de la prise en charge de ces élèves, les résultats du présent article montrent que les dispositifs orthopédagogiques de Programme Educatif Individualisé mis en place dans cette école ont produit de bons résultats en termes d’évolution ou d’amélioration de l’état des cas des élèves analysés. Cet accompagnement desdits élèves se fait en fonction de type de handicap mental et des capacités réelles de ces élèves. La présence du personnel qualifié et de matériels ainsi que le milieu d’accueil adapté ont favorisé la meilleure prise en charge de ces élèves. Cela montre que le Village Bondeko a une bonne organisation pour une meilleure prise en charge des élèves en situation de handicap mental modère et sévère. Mots-clés : Orthopédagogie, prise en charge scolaire, handicap mental, Programme éducatif individualisé, école spéciale.

Antonio Di Marco, Roberto Navigli

Umar Shehzad

The present study proposes that Samuel Beckett's famous reluctance to engage in the textualization of his personal life may be rooted in his apprehension towards the potential constriction of his self in the medium of language. This apprehension may have been further exacerbated by the fear of his life becoming subsumed by the written word. The notion that "we are what we say we are" postulates that one's identity is shaped by their own self-expression. In the context of biography, this notion can be extended to "we are what is said we are." Explaining his concept of interpellation, Louis Althusser, posits that various ideological apparatuses create a set of assumptions, attitudes, and desires that constitute an individual's self-identity. I argue that biography being an ideological apparatus of its own kind might do the same. Beckett's reluctance to engage in the textualization of his personal life may be a result of his desire to avoid the possibility of his self-identity being interpellated or defined by such texts. This may also explain his attempts to eliminate recognizable markers of his life from his work and his assertion of an absolute disconnection between his life and his work.

Ludovica Lena

Availing ourselves of a parallel corpus – composed of two written and spoken subcorpora –, in this study we adopt a contrastive perspective to explore the (lack of) correspondence between French il y a-constructions with a definite pivot and Chinese yǒu-constructions. It is shown that most il y a-constructions appear in enumerative contexts (including the sentence-level and inter-clause list-reading, as well as the instantiation of an explicit part-whole relation). Written Chinese uses other strategies to ensure textual coherence and to mark a part-whole relation between the “pivot” and a given referential group (e.g., the lexical verb bāokuò “include”). In the spoken subcorpus, the yǒu-construction is the most frequent strategy which however competes with the locative shì “be” construction when a Ground is overtly expressed. Presentational and event-reporting il y a-constructions are mostly found in the spoken corpus and align more neatly with yǒu-constructions. Overall, the present study argues that the differences between the two languages are more nuanced than has so far been believed, showing that the correspondence between the cognitive status of referents and the formal marking of definiteness is not straightforward. Chinese pivots that are unmarked with respect to definiteness (e.g., bare nouns and bare-head noun phrases) can be semantically definite despite their postverbal position. At the same time, French definite pivots do not always denote identifiable referents, as is the case of the weak use of the definite article but also the special use of demonstrative determiners in topic-promoting and event-reporting contexts.

Alon Talmor, Yanai Elazar, Yoav Goldberg et al.

Anastasia Shimorina, Yannick Parmentier, Claire Gardent

AbstractThe metrics standardly used to evaluate Natural Language Generation (NLG) models, such as BLEU or METEOR, fail to provide information on which linguistic factors impact performance. Focusing on Surface Realization (SR), the task of converting an unordered dependency tree into a well-formed sentence, we propose a framework for error analysis which permits identifying which features of the input affect the models’ results. This framework consists of two main components: (i) correlation analyses between a wide range of syntactic metrics and standard performance metrics and (ii) a set of techniques to automatically identify syntactic constructs that often co-occur with low performance scores. We demonstrate the advantages of our framework by performing error analysis on the results of 174 system runs submitted to the Multilingual SR shared tasks; we show that dependency edge accuracy correlate with automatic metrics thereby providing a more interpretable basis for evaluation; and we suggest ways in which our framework could be used to improve models and data. The framework is available in the form of a toolkit which can be used both by campaign organizers to provide detailed, linguistically interpretable feedback on the state of the art in multilingual SR, and by individual researchers to improve models and datasets.1

Andrea Moro, Alessandro Raganato, Roberto Navigli

Zhiguo Wang, Chengqing Zong

Driss LOUIZ

A l’heure de la pandémie du  Coronavirus  et  après  l’état d’urgence  sanitaire déclaré par le royaume du Maroc,  le Ministère de l’Education Nationale de la Formation Professionnelle, de l’Enseignement Supérieur et de la Recherche  Scientifique a pris d’importantes mesures préventives afin d’assurer  la continuité pédagogique de l’année 2019/2020 et  a décidé la suspension des cours  dans  les  différents  établissements scolaires  et universitaires des secteurs publics  et privés. Le   but est, en effet,  de lutter   contre  la propagation  de la pandémie  Covid-19 à partir  du 16  mars,  et  ce jusqu’à 10 juillet 2020 .  Les cours en présentiel ont  été donc remplacés  par   un enseignement à distance. Dans cet article nous essayerons, d’abord,  d’aborder la question de l’enseignement à distance à l’université ibn Tofail au temps du Covid-19, ensuite nous décrirons brièvement l’état des lieux   de l’enseignement/apprentissage  à l’heure  de l’épidémie  du Corona virus puis, nous mettrons l’accent sur la démarche  suivie par l’université  pour assurer  la continuité pédagogique par la mise en place des différentes plateformes de formation comme Google Classroom  et  enfin,  cet  article  s’interrogera  sur les   leçons à tirer  de  cette  pandémie. La problématique majeure  afférente  à cette recherche est de montrer comment  faire  bon usage des technologies de l’information et de la communication dans un contexte à la fois difficile  et complexe  pour assurer assurer le suivi pédagogique durant cette   situation de crise inopinée. Notre  recherche  s’inscrit  dans des débats théoriques tels  que  le  socio-constructivsime  combiné à la perspective  actionnelle  qui  considère l’apprenant comme  un acteur social et lui accorde un rôle central dans  son apprentissage. .

Anjar Dwi Astuti

African literature has strong relation with colonialism, not only because they had ever been colonized but also because of civil war. Civil Peace (1971), a short story written by Chinua Achebe, tells about how Nigerian survive and have to struggle to live after Nigerian Civil War. It is about the effects of the war on the people, and the “civil peace” that followed. The Nigerian Civil War, also known as the Nigerian-Biafran War, 6 July 1967–15 January 1970, was a political conflict caused by the attempted annexation of the southeastern provinces of Nigeria as the self-proclaimed Republic of Biafra. The conflict was the result of economic, ethnic, cultural and religious tensions among the various peoples of Nigeria. Knowing the relation between the story and the Nigerian Civil War, it is assured that there is a history depicted in Civil Peace. In this article, the writer portrays the history and the phenomenon of colonization in Nigeria by using new historical and postcolonial criticism approaches. Keywords: history, colonization, civil war

Markus Rheindorf

Academic discourse relies on a common register that reaches across disciplinary boundaries. This so-called common language of academia includes terms denoting common objects and actions of academic work in the fields of knowing, understanding and learning, of perceiving and observing, of designing and devising as well as of writing and speaking in their many different forms. In German academic discourse, this common register is highly figurative, often creating problems in bilingual contexts as well as during students’ acquisition of academic discourse. This paper presents an ongoing project to identify, describe and categorize the lexical dimension of this common language of academia by means of corpus-linguistic analyses. It presents operators to distinguish said register from non-academic German as well as discipline-specific discourse. Following these filtering procedures, the figurativity of the identified items is discussed in terms of their spatial-haptic, optical-visual, mental-cognitive and acoustic-linguistic character. Prominent cases of ambiguity, when specific lexical items can be used and read as expressing two figurative aspects, are also identified.

Antonio Perri

In un saggio scritto più di trent’anni or sono e ripubblicato dopo la sua morte Giorgio Raimondo Cardona si sforzava di valutare l’utilità di una teoria grafemica struttura- lista modellata sulla fonologia, e ricorrendo ad esempi tratti da sistemi di scrittura noti e meno noti (ma sempre di tipo fonetico) metteva in luce la scarsa applicabilità di una nozione di grafema che “non è altro che un nome dotto per ‘lettera dell’alfabeto’”. Cardona sottolineava tre gravi limiti della “grafemica autonoma”: • 1. Poiché l’isomorfismo tra catena parlata e catena scritta non è sempre perfetto, la grafemica non spiega fenomeni come le abbreviazioni grafiche o ad es. i “complessi grafici” del tibeta- no. • 2. La individuazione delle “unità minime” su basi fonologiche (i grafemi) – e, negli ulteriori sviluppi della teoria, quella dei tratti distintivi grafici – è frutto di un pregiudizio “tipografi- co” e non trova riscontro effettivo nei criteri di elaborazione di particolari sistemi grafici, come l’armeno o il palhavi. • 3. La grafemica autonoma non è suscettibile di ampliamenti che affrontino la costellazione di fatti sociologici connessi all’uso di un sistema di scrittura. Si tratta di verificare l’efficacia delle brillanti proposte che il compianto linguista aveva avanza- to per superare i limiti della grafemica in vista di una teoria “integrata” della scrittura.

Chantal Rittaud-Hutinet

To what extent non-recorded oral corpora may constitute objects of analysis of pragmatic meaning? These corpora are heard by chance: on the radio, on television, in the street, a shop, a means of transport or generally in any conversational interaction in which the linguist participates, but had not previously planned to record for his research. The problem of the use of these corpora in linguistics is all the more crucial since the aim, in phonopragmatics, is to discover the functions and significations of their phonic part. I shall attempt to answer the following questions: – The accuracy of the transcription with respect to the original. To what extent can we ignore our own phonological code, our regional variants, mastered/partly known styles of speech? – The reliability of the oral reproduction carried out by the linguist – for example, during a talk at a conference. What is his capacity for deferred mimicry? – The relation between a significant discrepancy and the elocutionary habits of the speaker. – The relation between the comprehension of the external auditors and the effect produced on the 'real' person addressed. Considering that transparency is (sometimes? often?) an illusion, I shall also examine what precautions should be taken so that these corpora offer guarantees as to the veracity.

Benke, Gertraud

In this article, the status of applied linguistics as discipline is questioned and problems of establishing it - and other newly formed scientific enterprises like cultural science - as disciplines are discussed. This discussion is contextualized using the author's own experience as applied linguist working in (the institutional structure of) Austria. Secondly, applied linguistics is presented as complementing cultural science, with both exploring at times the same phenomena albeit under different perspectives and focussing on different levels of experience. Two examples of research involving such a joint interest with different foci are discussed.