Basée sur une analyse contrastive, cette thèse aborde la structure verbale en français et en russe. Son objectif est d’établir un regard réflexif sur les deux langues dans le but de déterminer la différence et la ressemblance dans leur construction verbale. Explorant la question des verbes de structure différente des langues concernées, elle tente également de trouver les difficultés provoquées par cette différence à l’apprentissage. Ainsi, dans cette recherche, on parle des verbes problématiques sous un angle aussi didactique que linguistique. Composé de trois chapitres, ce travail étudie dans un premier temps le statut transitif/intransitif des verbes dans les deux langues en traitant l’approche traditionnelle et moderne et met en place une étude contrastive par rapport à la question de valence. On explore la construction verbale avec complément dans le deuxième chapitre de la recherche. Dans cette partie, l’analyse parallèle des verbes est effectuée afin de repérer leur fonctionnement selon les moyens grammaticaux de chaque langue. Et finalement, la comparaison de chaque verbe considéré problématique à l’apprentissage des deux langues suivie de tableaux fait partie du dernier chapitre. La production des tests préliminaires effectués auprès des apprenants russophones constitue également cette partie pour mieux comprendre la difficulté de ces derniers lors de la communication. / Based on a contrastive analysis, this thesis deals with the verbal structure in French and Russian. Its objective is to establish a reflexive look at the two languages in order to determine the difference and the similarity in their verbal construction. Exploring the question of the verbs of different structure, it also tries to find the difficulties caused by this difference in learning. Thus, in this research, problematic verbs are spoken of in a didactic as well as linguistic angle. This work, composed of three chapters, speaks first of all about the transitive/intransitive status of verbs in both languages by treating the traditional and modern approach and sets up a contrastive study in relation to the valence question. Verbal construction is explored with complement in the second chapter of the research. In this part, the parallel analysis of the verbs is performed in order to identify their functioning according to the grammatical means of each language. And finally, the comparison of each verb considered problematical to the learning in both languages followed by tables took the part of the last chapter. The production of preliminary tests with Russian-speaking learners is also part of this work in order to better understand the difficulties of the latter during the communication.

Nesta dissertação, foi desenvolvida uma análise linguística contrastiva de bulas de medicamento brasileiras e alemãs. O principal objetivo dessa análise era identificar o grau de compreensibilidade de cada um dos textos, segundo as dimensões do modelo de Karlsruhe: Concisão, Correção, Motivação, Estrutura, Simplicidade e Perceptibilidade. Para isso, foram retomados conceitos da Linguística Textual e procedimentos da Textologia Contrastiva, como as definições de texto e gênero textual, o método de comparação de textos paralelos e o estabelecimento de um termo de comparação. De maneira geral, os resultados da análise contrastiva indicaram que termos técnicos (no âmbito da simplicidade) e fonte pequena (no âmbito da perceptibilidade) não são as principais causas de incompreensão na leitura das bulas, como se supôs inicialmente. O estudo evidenciou que desvios na estrutura e na concisão das bulas podem ter um impacto tão ou mais negativo na compreensão desses textos do que os aspectos anteriormente citados. Além disso, foi possível constatar que cada uma das dimensões possui um peso diferente na compreensibilidade das bulas analisadas devido a diferenças linguísticas e culturais entre os pares alemão/português e Alemanha/Brasil. / In this research, a contrastive linguistic analysis of German and Brazilian package inserts was developed. The main aim of this analysis was to identify the degree of comprehensibility in each text, according to the Karlsruhe model dimensions: concision, correctness, motivation, structure, simplicity and perceptibility. To achieve this aim, it was necessary to review some concepts of the Textual Linguistics and procedures of the Contrastive Textology, such as the definitions of text and textual genre, the method of comparing parallel texts and the establishment of a point of comparison. In general, the results of the contrastive analysis have shown that technical terms (in the scope of simplicity) and small font (perceptibility) are not the main causes of misunderstanding of package inserts, as it was supposed initially. The study has highlighted that deviances in the structure and in the concision of package inserts may have an impact in the comprehension of these texts as negative as the aspects forementioned, if not more. Moreover, it was possible to notice that each dimension has a different weight in the comprehensibility of the analyzed package inserts because of the linguistic and cultural differences between German/Portuguese and Germany/Brazil.

The subject of this diploma thesis is the use of Turkisms in Bulgarian, Greek and Albanian languages. First of all, the basic principles of lexical borrowing and the place of Turkisms in the current linguistic situation of the Bulgarian, Greek and Albanian languages are discussed. Attempts have been made to identify a specific number of Turkisms in each language, and some views have been presented on the term "Turkism", according to some scholars. Periodization of the onset of Turkisms are discussed, as well as a brief overview of previous studies on Turkisms as an introduction to the problem. After that, the subject of the work went on to the morphological and lexical-semantic analysis of Turkisms. From a morphological point of view, the basic Turkish suffixes borrowed in these three Balkan languages and their main meanings are considered. The analysis is performed with the selection of 70 Turkisms from the index of Prof. dr. Maxim Stamenov and having as source language for the comparison the Bulgarian language. The most famous and extensive dictionaries of each language from which the meanings of the lemmas came from were selected for comparison. In the same way, the analysis of Bulgarian, Greek and Albanian Turkisms in the language of the media is carried out. I have tried to find examples of...

Since the advent of automatic evaluation, tasks within Natural Language Processing (NLP), including Machine Translation, have been able to better utilize both time and labor resources. Later, multilingual pre-trained models (MLMs)have uplifted many languages’ capacity to participate in NLP research. Contextualized representations generated from these MLMs are both influential towards several downstream tasks and have inspired practitioners to better make sense of them. We propose the adoption of BERTScore, coupled with contrastive learning, for machine translation evaluation in lieu of BLEU - the industry leading metric. While BERTScore computes a similarity score for each token in a candidate and reference sentence, it does away with exact matches in favor of computing token similarity using contextual embeddings. We improve BERTScore via contrastive learning-based fine-tuning on MLMs. We use contrastive learning to improve BERTScore across different language pairs in both high and low resource settings (English-Hausa, English-Chinese), across three models (XLM-R, mBERT, and LaBSE) and across three domains (news,religious, combined). We also investigated both the effects of pairing relatively linguistically similar low-resource languages (Somali-Hausa), and data size on BERTScore and the corresponding Pearson correlation to human judgments. We found that reducing the distance between cross-lingual embeddings via contrastive learning leads to BERTScore having a substantially greater correlation to system-level human evaluation than BLEU for mBERT and LaBSE in all language pairs in multiple domains.

No description available.

The knowledge of tests conducted on a problematic vehicle is essential for enhancing the efficiency of mechanics. Therefore, identifying the tests performed in each repair case is of utmost importance. This thesis explores techniques for extracting data from unstructured repair reports to identify component tests. The main emphasis is on developing a supervised multi-class classifier to categorize data and extract sentences that describe repair diagnoses and actions. It has been shown that incorporating a category-aware contrastive learning objective can improve the repair report classifier’s performance. The proposed approach involves training a sentence representation model based on a pre-trained model using a category-aware contrastive learning objective. Subsequently, the sentence representation model is further trained on the classification task using a loss function that combines the cross-entropy and supervised contrastive learning losses. By applying this method, the macro F1-score on the test set is increased from 90.45 to 90.73. The attempt to enhance the performance of the repair report classifier using a noisy data classifier proves unsuccessful. The noisy data classifier is trained using a prompt-based fine-tuning method, incorporating open-ended questions and two examples in the prompt. This approach achieves an F1-score of 91.09 and the resulting repair report classification datasets are found easier to classify. However, they do not contribute to an improvement in the repair report classifier’s performance. Ultimately, the repair report classifier is utilized to aid in creating the input necessary for identifying component tests. An information retrieval method is used to conduct the test identification. The incorporation of this classifier and the existing labels when creating queries leads to an improvement in the mean average precision at the top 3, 5, and 10 positions by 0.62, 0.81, and 0.35, respectively, although with a slight decrease of 0.14 at the top 1 position.

In the banking world trades of securities are finalized every day, on behalf of the banks themselves or of their clients. When the trades have been booked by the front office the confirmations sent by the counterparty have to be checked and connected to the correct trade by hand, posing the question whether this process could not be automated using machine learning techniques. There is no straightforward solution to this problem since the confirmations differ between counterparties, and can contain different enriched information or even be in different formats. This thesis addresses the problem of matching trades with their corresponding confirmations via deep learning methods. A model is trained using contrastive learning methods on generated pairs of trades and confirmations, with the goal of matching the pairs in the latent space by using nearest neighbor classification. Accuracy is measured by dividing the correctly classified samples by the total number of samples in a testing batch. The model achieves an accuracy as high as 97.8% over 100 trade-confirmation samples with a 30-dimensional latent space, and it is shown that similar contrastive methods can indeed be used in order to solve this problem. / Banker handlar varje dag med värdpapper av olika slag, antingen för sin egen vinning eller för sina kunders. När en affär har blivit beslutad mellan två parter så bokförs denna i bägge parternas interna system. En konfirmation kommer sedan skickas från den andra parten som manuellt måste paras ihop med affären vilket väcker frågan om huruvida detta inte kan automatiseras med hjälp av maskininlärning. Det finns inte en uppenbar lösning på detta problemet då konfirmationsmeddelandena kan skiljer sig åt mellan olika parter och kan innehålla olika tillagd information eller till och med vara i olika format. En model tränas genom att använda kontrast-inlärning på genererade par av affärer och konfirmationer av affärer för att kunna para ihop paren i det latenta rummet genom att se vilka grannar som ligger närmast. Nogrannheten mäts genom att dela antalet korrekt klassificerade exempel med det totala antalet par i en grupp test-par. Modellen uppnår en noggrannhet så hög som 97.8% på 100 affärs-konfirmationspar med ett 30-dimensionellt latent rum, och det visas att kontrast-inlärning kan användas för att lösa problemet. Det är dock svårt att säga mycket om hur väl modellen kan generalisera de inlärda kunskaperna eftersom träningsdatan behövde genereras och därför saknar en del av komplexiteten av ett äkte data set.

The present study is based on a translation into Swedish of three French journalistic articles on energy sobriety. The study aims at comparing the source and the target texts, with regard to clause frequency and the use of various clause and phrase categories. This analysis is linked to the structural differences between the two languages, that have been stated by researchers in contrastive linguistics. The clause and phrase categories in focus have been defined by Olof Eriksson (1997), and used as translation units in his contrastive study on the French and Swedish languages.  The results of the analysis prove that the Swedish translation is characterised by a considerably higher clause frequency than the French source texts. The reason is that the translation has led to more phrases being replaced by clauses, than the opposite. This divergency is conform to the stated differences between French and Swedish – a conclusion that also can be drawn regarding the most frequent shifts from phrase to clause categories, especially the ones including a participle phrase. However, not all observed shifts are obligatory, which means that the results could be somewhat different if the analysis was based on a translation conducted by someone else. In a future study, it would therefore be interesting to extend the analysis to include several versions of the same translated text.

Surface topography systems enable the capture of spinal dynamic movement. A visualization of possible unique movement patterns appears to be difficult due to large intraclass and small inter-class variabilities. Therefore, we investigated a visualization approach using Siamese neural networks (SNN) and checked, if the identification of individuals is possible based on dynamic spinal data. The presented visualization approach seems promising in visualizing subjects in the presence of intraindividual variability between different gait cycles as well as day-to-day variability. Overall, the results indicate a possible existence of a personal spinal ‘fingerprint’. The work forms the basis for an objective comparison of subjects and the transfer of the method to clinical use cases.

Unnormalised parametric models are an important class of probabilistic models which are difficult to estimate. The models are important since they occur in many different areas of application, e.g. in modelling of natural images, natural language and associative memory. However, standard maximum likelihood estimation is not applicable to unnormalised models, so alternative methods are required. Noise-contrastive estimation (NCE) has been proposed as an effective estimation method for unnormalised models. The basic idea is to transform the unsupervised estimation problem into a supervised classification problem. The parameters of the unnormalised model are learned by training the model to differentiate the given data samples from generated noise samples. However, the choice of the noise distribution has been left open to the user, and as the performance of the estimation may be sensitive to this choice, it is desirable for it to be automated. In this thesis, the ambiguity in the choice of the noise distribution is addressed by presenting the previously unpublished conditional noise-contrastive estimation (CNCE) method. Like NCE, CNCE estimates unnormalised models by classifying data and noise samples. However, the choice of noise distribution is partly automated via the use of a conditional noise distribution that is dependent on the data. In addition to introducing the core theory for CNCE, the method is empirically validated on data and models where the ground truth is known. Furthermore, CNCE is applied to natural image data to show its applicability in a realistic application. / Icke-normaliserade parametriska modeller utgör en viktig klass av svåruppskattade statistiska modeller. Dessa modeller är viktiga eftersom de uppträder inom många olika tillämpningsområden, t.ex. vid modellering av bilder, tal och skrift och associativt minne. Dessa modeller är svåruppskattade eftersom den vanliga maximum likelihood-metoden inte är tillämpbar på icke-normaliserade modeller. Noise-contrastive estimation (NCE) har föreslagits som en effektiv metod för uppskattning av icke-normaliserade modeller. Grundidén är att transformera det icke-handledda uppskattningsproblemet till ett handlett klassificeringsproblem. Den icke-normaliserade modellens parametrar blir inlärda genom att träna modellen på att skilja det givna dataprovet från ett genererat brusprov. Dock har valet av brusdistribution lämnats öppet för användaren. Eftersom uppskattningens prestanda är känslig gentemot det här valet är det önskvärt att få det automatiserat. I det här examensarbetet behandlas valet av brusdistribution genom att presentera den tidigare opublicerade metoden conditional noise-contrastive estimation (CNCE). Liksom NCE uppskattar CNCE icke-normaliserade modeller via klassificering av data- och brusprov. I det här fallet är emellertid brusdistributionen delvis automatiserad genom att använda en betingad brusdistribution som är beroende på dataprovet. Förutom att introducera kärnteorin för CNCE valideras även metoden med hjälp av data och modeller vars genererande parametrar är kända. Vidare appliceras CNCE på bilddata för att demonstrera dess tillämpbarhet.