Hasil untuk "Slavic languages. Baltic languages. Albanian languages"

Jānis Taurēns

Recenzija par grāmatu: Klusumā. Kvīri, padomju vara un sabiedrība Latvijā, 1954–1991. Ineta Lipša, Kārlis Vērdiņš un Kaspars Zellis (red.). Rīga 2024. 351 lpp. ISBN 978-9934-615-20-7

Taja Kuzman Pungeršek, Peter Rupnik, Ivan Porupski et al.

Until recently, fine-tuned BERT-like models provided state-of-the-art performance on text classification tasks. With the rise of instruction-tuned decoder-only models, commonly known as large language models (LLMs), the field has increasingly moved toward zero-shot and few-shot prompting. However, the performance of LLMs on text classification, particularly on less-resourced languages, remains under-explored. In this paper, we evaluate the performance of current language models on text classification tasks across several South Slavic languages. We compare openly available fine-tuned BERT-like models with a selection of open-source and closed-source LLMs across three tasks in three domains: sentiment classification in parliamentary speeches, topic classification in news articles and parliamentary speeches, and genre identification in web texts. Our results show that LLMs demonstrate strong zero-shot performance, often matching or surpassing fine-tuned BERT-like models. Moreover, when used in a zero-shot setup, LLMs perform comparably in South Slavic languages and English. However, we also point out key drawbacks of LLMs, including less predictable outputs, significantly slower inference, and higher computational costs. Due to these limitations, fine-tuned BERT-like models remain a more practical choice for large-scale automatic text annotation.

Yifan Zong, Yuntian Deng, Pengyu Nie

Large language models (LLMs) have demonstrated impressive capabilities in aiding developers with tasks like code comprehension, generation, and translation. Supporting multilingual programming -- i.e., coding tasks across multiple programming languages -- typically requires either (1) finetuning a single LLM across all programming languages, which is cost-efficient but sacrifices language-specific specialization and performance, or (2) finetuning separate LLMs for each programming language, which allows for specialization but is computationally expensive and storage-intensive due to the duplication of parameters. This paper introduces MoLE (Mix-of-Language-Experts), a novel architecture that balances efficiency and specialization for multilingual programming. MoLE is composed of a base model, a shared LoRA (low-rank adaptation) module, and a collection of language-specific LoRA modules. These modules are jointly optimized during the finetuning process, enabling effective knowledge sharing and specialization across programming languages. During inference, MoLE automatically routes to the language-specific LoRA module corresponding to the programming language of the code token being generated. Our experiments demonstrate that MoLE achieves greater parameter efficiency compared to training separate language-specific LoRAs, while outperforming a single shared LLM finetuned for all programming languages in terms of accuracy.

Majk Johansen

Italian translation of Elementarni zakony versyfikatsii by Maik Iohansen.

Elena Jukhimenko

The article highlights the work which led to the publication in 2010 of the Russian translation of Pierre Pascal’s famous study The Archpriest Avvakum and the Beginning of the Schism. This edition was based on the translation commissioned by the Moscow Old Believer Archdiocese from the grandson of writer Leo Tolstoy, Sergei Sergeevich Tolstoy, from the late 1940s to the mid-1950s. Since the translation was initially not intended for printing, it had to undergo considerable scientific editing when it was published in 2010. This included clarifying bibliographical mentions, checking titles of Old-Russian works, as well as quotations. The two consecutive Moscow editions of the French scholar’s monograph, which came out in 2010 and 2016 respectively, testify to the enduring significance of this fundamental work.

Sugai, Kenta

Based on quantitative and contrastive analyses, this study explores the borrowing of verbs in the Bulgarian dialect of Brăneşti in Romania, how bilingual speakers integrate Romanian loan verbs, and why such integration strategies are adopted. For this purpose, the original colloquial data collected by the author in the 2010s, which included 93 Romanian loan verbs, are used. I argue that bilingual speakers choose either a direct or indirect insertion strategy, and in the latter case, they utilise the borrowed morpheme -’askă as an integration marker. The process is motivated by formal similarities and productivity.

Thor Bjørn Arlov

At the outbreak of the First World War there was virtually no Norwegian coalmining activity on the Spitsbergen archipelago. The handful of small coal companies that were formed in Norway around the turn of the century were either idle or had been bought up by foreign interests after a few years. During the war, however, several new private companies were established, most notably the Store Norske Spitsbergen Kulkompani in 1916. Two years later, in 1918–1919, the Norwegian government stated its desire to acquire full sovereignty over the archipelago. The wish was granted by the treaty of 9 February 1920 that came as a result of the peace negotiations in Paris. This paper reviews the role of the Norwegian coal companies in Norway’s quest for supremacy over Svalbard during and after the First World War. Were private enterprises an instrument of the Norwegian government’s ambitions or was it the other way around? It is argued that private companies were instrumental in moving the political authorities from a passive to an active stance regarding sovereignty during the last phase of the war and through the peace conference in 1919. Their primary concern was to protect their own vested interests. However, as soon as sovereignty was secured in 1920, it was the government that actively used the companies as instruments to improve Norway’s position on the archipelago before implementing the treaty and settling the property rights. Note: I use the official toponym ‘Svalbard’, although before 1925 ‘Spitsbergen’ was more commonly used.

Dzhonova, Marina

Syntactic Description of Mental-State Predicative Constructions. The paper examines the argument structure of constructions containing mental-state predicative words in the contemporary Bulgarian language. The source material was extracted from the Bulgarian National Corpus. The analysis focuses on predicates denoting knowledge, confidence and doubt. The research reveals that the role of such predicates can be performed by predicative adverbs (‘yasno’, ‘izvestno’, ‘samnitelno’), predicative nouns (‘istina’, ‘fact’) or predicative adjectives (‘uveren’, ‘siguren’). Cases containing explicit or implicit experiencer arguments are discussed along with the possibilities for expressing the object and the content of knowledge. Some adverbs could express the experiencer’s degree of confidence. The analysis of the empirical data leads to conclusions about the common argument structure of predicative constructions denoting mental state and for the syntactic expression of the arguments.

Badr M. Abdullah, Tania Avgustinova, Bernd Möbius et al.

State-of-the-art spoken language identification (LID) systems, which are based on end-to-end deep neural networks, have shown remarkable success not only in discriminating between distant languages but also between closely-related languages or even different spoken varieties of the same language. However, it is still unclear to what extent neural LID models generalize to speech samples with different acoustic conditions due to domain shift. In this paper, we present a set of experiments to investigate the impact of domain mismatch on the performance of neural LID systems for a subset of six Slavic languages across two domains (read speech and radio broadcast) and examine two low-level signal descriptors (spectral and cepstral features) for this task. Our experiments show that (1) out-of-domain speech samples severely hinder the performance of neural LID models, and (2) while both spectral and cepstral features show comparable performance within-domain, spectral features show more robustness under domain mismatch. Moreover, we apply unsupervised domain adaptation to minimize the discrepancy between the two domains in our study. We achieve relative accuracy improvements that range from 9% to 77% depending on the diversity of acoustic conditions in the source domain.

Krasimir Yordzhev

Context-free languages are widely used to describe the syntax of programming languages and natural languages. Usually, we describe a context-free language mathematically with the help of context-free grammar (for generation) or pushdown automata (for recognition). The purpose of this study is to describe some unconventional methods of description of context-free languages, namely a representation with the help of finite digraphs and with automata - generators of context-free languages. We will mainly focus on the mathematical models of these representations.

Caterina Corbella

Aim of this paper is to detect the principles that guided the Catholic theologist Romano Guardini when working with literature, using as an example his works on Dante and Dostoevsky. The paper is divided into three parts. The first part identifies some key moments that Guardini shares with Dostoevsky and Dante about man and the world, most notably about the process of understanding. The reader is thus given not only the context of Guardini’s though, but also a suggestion about the reason of the closeness he felt with those authors. The second part of the paper is dedicated to Guardini’s experience as professor of Catholic Weltanschauung in Berlin, as the course he held there became a sort of breeding ground where his method of reading literature was born. The last part of the paper focuses on Guardini’s essays about Dante and Dostoevsky, and through concrete examples discloses the guidelines of his works under the light of the previous parts.

Peter D. Mosses

The CBS framework supports component-based specification of programming languages. It aims to significantly reduce the effort of formal language specification, and thereby encourage language developers to exploit formal semantics more widely. CBS provides an extensive library of reusable language specification components, facilitating co-evolution of languages and their specifications. After introducing CBS and its formal definition, this short paper reports work in progress on generating an IDE for CBS from the definition. It also considers the possibility of supporting component-based language specification in other formal language workbenches.

Oscar H. Ibarra, Ian McQuillan, Bala Ravikumar

We study counting-regular languages -- these are languages $L$ for which there is a regular language $L'$ such that the number of strings of length $n$ in $L$ and $L'$ are the same for all $n$. We show that the languages accepted by unambiguous nondeterministic Turing machines with a one-way read-only input tape and a reversal-bounded worktape are counting-regular. Many one-way acceptors are a special case of this model, such as reversal-bounded deterministic pushdown automata, reversal-bounded deterministic queue automata, and many others, and therefore all languages accepted by these models are counting-regular. This result is the best possible in the sense that the claim does not hold for either $2$-ambiguous PDA's, unambiguous PDA's with no reversal-bound, and other models. We also study closure properties of counting-regular languages, and we study decidability problems in regards to counting-regularity. For example, it is shown that the counting-regularity of even some restricted subclasses of PDA's is undecidable. Lastly, $k$-slender languages -- where there are at most $k$ words of any length -- are also studied. Amongst other results, it is shown that it is decidable whether a language in any semilinear full trio is $k$-slender.

E. N. Yemelyanova

The Bolshevik policy toward Germany, the relationship between the Communist Party of Soviet Russia and the German Social Democracy, which came to power as a result of the First World War is analyzed in the article. The issue of the geopolitical interests and ideological orientations of the Bolshevik leadership and the German socialists is considered, and disagreements between various currents in Marxism on internal and foreign policy issues are examined. The problem of the dualism of the policy of the CPSU (b) in the first years of power, the contradiction between the theoretical attitudes to the "world revolution" and the practical need of the Soviet state for carrying out a "policy of peaceful coexistence" is raised, and the strategy and tactics of the Bolsheviks are analyzed at a time when their government turned out the emerging Versailles system of international relations. The activity of the Communist International on the expansion of political influence in the world and the search for allies for Soviet Russia are traced. The topicality of the problem under investigation is conditioned by the need to develop long-standing historical ties between the Russian and German states, to study the mechanisms for establishing good-neighborly relations and the desirability of improving them at the present stage. The author concludes that, despite theoretical differences, the national interests of Germany and Russia forced both Marxist parties to converge, which ultimately contributed to the conclusion of the Rapallo Treaty and the transition to political and economic cooperation between the two countries.

E. A. Kotelenets, M. S. Zatulovskaya, M. Y. Lavrentieva

The process of forming and changing the image of Russia-bear is analyzed. English, French, Polish, American and Georgian caricatures, engravings and lithographs, which appeared in the period from the second half of the 18th century to the present day, are used as sources. The problem of forming the image of Russia abroad is considered from the point of view of imagology and information warfare. The process of changing the image of Russia-bear is described depending on the political situation, on the direction and propagandistic goals of the printed publications. The study revealed that the associations of Russia with the image of the bear are extremely strong and acquired the status of a stereotype long ago. It is noted that this image has been successfully used as a weapon of political propaganda since the XVIII century. It is shown that the image of the "Russian bear" in the Western media constantly changes on the basis of the current political situation on the international arena. The authors assert that this image is practically not developed in the Western media in the context of constructive alignment of international relations, the destructive component of the image is usually emphasized. It is reported that, since 1977, attempts have been made in Russia to positivize the image of the bear in foreign media, but they have not been successful to the present day.

Dalia Čiočytė

[straipsnis lietuvių kalba; santrauka anglų kalba] Straipsnyje tyrinėjama šiuolaikinėje lietuvių eseistikoje matoma transcendencijos samprata ir jos poveikis literatūrinio aš savimonei. Derinant teorinę literatūros teologijos perspektyvą su Michailo Bachtino paslėptos polemikos samprata, analizuojama pasakotojo, autoriaus atvira ir slapta polemika su metafiziniu adresatu trijuose reprezentaciniuose eseistikos tekstuose: Sigito Parulskio „Aukojime“ (2002), Giedros Radvilavičiūtės „Teksto traukoje“ (2010), Alfonso Andriuškevičiaus „Atsisveikinime su daiktais“ (2015). Šiuose daugiabalsiuose tekstuose egzistencinis (egzistencinei savimonei lemtingai reikšmingas) dialogas įtaigiai deautomatizuoja žmogaus prasmės refleksiją. Daroma išvada, kad S. Parulskio „Aukojime“ vyrauja ontologinis literatūrinis mąstymas, postuluojantis žmogaus ir būties esmę, o G. Radvilavičiūtės „Teksto traukoje“ ir A. Andriuškevičiaus „Atsisveikinime su daiktais“ – gnoseologinis mąstymas, plėtojamas kaip intensyvi atsakymų į būties esmės ir žmogaus prasmės klausimus paieška.

Primož Vitez

Cankarjeva literatura je v svoji zvrstni in tematski raznolikosti dovolj kompleksna, da ne dopušča poenostavljenih analiz glede na tradicionalne literarno-teoretične in literarno-zgodovinske delitve. Čeprav je večino svojih besedil Ivan Cankar zapisal v prozi, je njegov slog v neprekinjenem stiku s poetičnim izvorom njegovega opusa. To ne velja samo za tekste, ki smo jih navajeni brati kot »čisto literaturo«, temveč tudi za tista besedila, s katerimi je Cankar dejavno posegal v kulturno politiko slovenstva v okvirih razpadajoče avstro-ogrske države. Enako kompleksen je njegov idejni pristop k obravnavanim političnim in bivanjskim tematikam, ki ga ne moremo učinkovito razčleniti s pristranskimi ideološkimi pristopi. Cankarjev socializem je prav tako uporniški kot njegov krščanski idealizem. Obe etični načeli, na katerih enakovredno temelji vse Cankarjevo pisanje, sta izpeljani in neločljivo povezani z izvirnimi estetskimi jezikovnimi prijemi, s katerimi je Cankar oplemenitil slovensko besedno umetnost ter »čistil in pomlajeval« politično mišljenje na Slovenskem.

Birzhan Moldagaliyev

We define a notion of randomness for individual and collections of formal languages based on automatic martingales acting on sequences of words from some underlying domain. An automatic martingale bets if the incoming word belongs to the target language or not. Then randomness of both single languages and collections of languages is defined as a failure of automatic martingale to gain an unbounded capital by betting on the target language according to an incoming sequence of words, or a text. The randomness of formal languages turned out to be heavily dependent on the text. For very general classes of texts, any nonregular language happens to be random when considered individually. As for collections of languages, very general classes of texts permits nonrandomness of automatic families of languages only. On the other hand, an arbitrary computable language is be shown to be nonrandom under certain dynamic texts.

Alexis Linard, Colin de la Higuera, Frits Vaandrager

A classical problem in grammatical inference is to identify a language from a set of examples. In this paper, we address the problem of identifying a union of languages from examples that belong to several different unknown languages. Indeed, decomposing a language into smaller pieces that are easier to represent should make learning easier than aiming for a too generalized language. In particular, we consider k-testable languages in the strict sense (k-TSS). These are defined by a set of allowed prefixes, infixes (sub-strings) and suffixes that words in the language may contain. We establish a Galois connection between the lattice of all languages over alphabet Σ, and the lattice of k-TSS languages over Σ. We also define a simple metric on k-TSS languages. The Galois connection and the metric allow us to derive an efficient algorithm to learn the union of k-TSS languages. We evaluate our algorithm on an industrial dataset and thus demonstrate the relevance of our approach.

A. V. Grigorovskaya

The results of the comparative analysis of two novels - by American author S. Lewis (“Arrowsmith”) and American writer of Russian origin Ayn Rand (“Altas shrugged”) - are presented. The images of scientists, the authors’ attitudes to the problems of science and money, science and power, science and reason, scientific and technological progress presented in the works are compared. The novelty of the research lies in the fact that these novels have never before been compared. Actuality of work is conditioned by the importance of the analysis of the different philosophical views of Lewis and Rand on socialism and capitalism, that found a contact in criticism of the contemporary political system of the first half of the 20th century. In the study the author notes that in both novels in different ways the position of science in a capitalist society, the relations of scientists and government representatives are shown, at that Lewis puts the consumer above the creator, and Rand lets the creator to be the center of the universe, towering above the crowd of consumers. It is alleged that the writers’ evaluation of the role of creators and consumers in the development of society can be projected in their attitude to the ideology of socialism and capitalism. However both authors warn about the evils of consumer attitudes to science in the contemporary political system.