?
A graph-based approach to closed-domain natural language generation
Научный результат. Серия: Вопросы теоретической и прикладной лингвистики. 2024. Vol. 10. No. 3. P. 135–167.
Victoria I. Firsanova
Graph-based Natural Language Processing (NLP) methods have seen significant advancements in recent years with the development of Large Language Models (LLMs) and Retrieval Augmented Generation (RAG). LLMs are sophisticated models that recognize numerous NLP tasks by analyzing the users' natural language instructions called prompts. However, their industrial use is questionable due to such ethical concerns as false information generation called hallucinations, high risks of data breaches, and plagiarism. The paper introduces a novel NLP architecture, the Graph-Based Block-to-Block Generation (G3BG), which leverages state-of-the-art deep learning techniques, the power of attention mechanisms, distributional semantics, graph-based information retrieval, and decentralized networks. The model encodes user prompts to mitigate data breach risk, retrieves relevant information from a graph knowledge base, and forms a block for a conditional language model using LLMs to perform a new secure type of RAG...
Keywords: Генеративный искусственный интеллектLarge Language ModelsLanguage GenerationLanguage UnderstandingGenerative Artificial IntelligenceDecentralized NetworksData EncodingDistributional SemanticsClosed-Domain SystemsГенерация естественного языкаПонимание естественного языкаБольшие языковые моделиДецентрализованные сетиКодирование данныхДистрибутивная семантикаЗакрытая предметная область
Vinogradov A., Rodionov O., Traveaux et memoires 2026 Vol. 30 No. 2 P. 1075–1089
This article examines the manuscript tradition of Kallistos Angelikoudes’ Hesychastic consolation, a fluid and complex corpus of thirty Logoi preserved primarily in Vat. gr. 736. The authors survey all known witnesses—including the autograph Barb. gr. 420, the Lond. Arund. 520, and several later copies—and analyze their textual relationships, variant readings, and codicological features. They demonstrate ...
Added: October 1, 2026
Громенко Е. С., Krongauz M., Антропологический форум 2026 № 70 С. 163–183
The paper describes the construction litso takoy-to natsionalnosti ‘a person of such-and-such an ethnicity’ in Russian with its semantic and pragmatic characteristics. The purpose of the study is to trace the appearance of the construction and analyse how it is used and how it is perceived. Three expressions representing this construction are considered in more ...
Added: September 30, 2026
Поцелуев В. А., Ural-Altaic Studies 2026 Vol. 62 No. 3 P. 62–81
In recent years, the researchers working on Uralic languages within Minimalism seem to have built a consensus on how to define the
position of negation in any language with a negative verb. The main and usually the only argument is based on the work [Mitchell 2006].
The argument is as follows: if the negative verb expresses tense, ...
Added: September 30, 2026
Государственный университет просвещения, 2026.
Настоящее издание представляет собой коллективный труд, посвящённый юбилеям основателей-руководителей научных школ лингвистического факультета Государственного университета просвещения Максименко Ольги Ивановны, доктора филологических наук, профессора, профессора кафедры теории языка, англистики и прикладной лингвистики, Председателя
диссертационного совета 72.2.020.05 и Хухуни Георгия Теймуразовича, доктора филологических
наук, профессора, профессора кафедры теории языка, англистики и прикладной лингвистики.
Цель издания – обобщение современных достижений в ...
Added: September 29, 2026
Zevakhina N., Щипкова А. А., Chinkova A., Voprosy Jazykoznanija 2025 No. 2 P. 105–122
This paper presents experimental evidence for the over-specification (redundant use) of Thai postnominal
numerals in referential communication. The first experiment reveals high rates of over-specification
in oral responses that include Thai postnominal and Russian prenominal numerals tested in a contrastive
multi-number visual context. The second experiment demonstrates lower but still relatively high
rates of over-specification in written responses that ...
Added: September 29, 2026
Вилинбахова Е. Л., Zevakhina N., Terra Linguistica 2025 Т. 16 № 3 С. 13–30
Настоящая работа посвящена исследованию употреблений испанского глагола resultar ‘оказываться’ в рамках эвиденциального подхода и подхода возникновения нового знания, представленных в литературе. Согласно первому подходу, глагол resultar передает
инферентивное значение: субъект оценки делает умозаключение на основании наблюдаемых
результатов. Сторонники второго подхода считают основным компонентом семантики глагола
возникновение нового знания, а инферентивное значение объясняется воздействием других
элементов контекста. В работе каждый ...
Added: September 29, 2026
Zevakhina N., Пономарева П., Краснощёкова С., Acta linguistica Petropolitana 2025 Vol. 21 No. 2 P. 270–286
The paper reports on an experimental study that reveals a difference
between two types of under-informative utterances with the quantifier “nekotoryj”
(‘some’) in Russian. The experiment was conducted as an inference task based on [Geurts,
Pouscoulous 2009; Val Tiel et al. 2016] with some modifications. The experiment
showed that both types of utterances more often have inference-based interpretations
rather than ...
Added: September 29, 2026
Zevakhina N., Voprosy Jazykoznanija 2026 No. 3 P. 54–70
The paper demonstrates that the Russian particle mol exhibits a combination of properties associated
with direct, indirect, and free indirect discourse (FID). First, mol-clauses license shiftable linguistic
items (adverbial indexicals, expressive speech acts, imperatives, evaluative lexical items) not
only in FID, as typically assumed for such linguistic items, but also in indirect discourse. Second,
first- and second-person pronouns in ...
Added: September 29, 2026
Sigdel E., Zevakhina N., Krasnoschekova S., Vestnik of Saint Petersburg University. Language and Literature 2026 Vol. 23 No. 1 P. 217–234
The paper investigates Russian interrogatives and exclamatives with multiple wh-phrases based on two experiments, with additional support of the data from the Russian National Corpus. Since exclamatives are often viewed parallel to interrogatives, it is important to study both types of clauses with multiple wh-phrases alongside and to look for similarities and/or differences in their ...
Added: September 29, 2026
Zevakhina N., Studi slavistici 2026 Vol. 23 No. 1 P. 111–126
Using syntactic embeddability diagnostics, the paper demonstrates that Russian wh-exclamatives
with the quotation particle mol and/or with the complementizer čto are embeddable and thus
contributes to the ongoing theoretical and cross-linguistic discussion of embedded wh-exclamatives
and their (in)subordination status. This analysis is applicable to rhetorical questions (with the quotation
particle mol and/or with the complementizer čto), which are traditionally ...
Added: September 29, 2026
Алентьева А. А., Новый филологический вестник 2026 Т. 77 № 2 С. 94–103
The article is devoted to research irony in Turgenev’s play A Month in the Country and Friel’s theatre adaptation A Month in the Country. The purpose of the study is to determine how the Irish playwright interprets Turgenev’s plot. Turgenev presents the moral decline of the Russian nobility amid a state of general disorder. Using ...
Added: September 28, 2026
Karpychev V., Malyutina S., Dragoy O., Brain Structure and Function 2026 Vol. 231 Article 69
Language comprehension and production are predominantly supported by the left-lateralized language-related network (LRN). However, right-hemispheric homologs of the language regions can also contribute to language processing, resulting in reduced left-hemispheric language lateralization. To date, it remains unclear whether less left-hemispheric language lateralization is associated with the recruitment of other non-language networks and their interaction with ...
Added: September 28, 2026
Аверьянова Варвара Андреевна, Sherbakova O., Вопросы психолингвистики 2022 Т. 53 № 3 С. 42–61
The article presents the results of the research aimed at building the framework for in-depth analysis of comprehension of implicit meanings in verbal texts. The existing approaches to studying implicit meanings and text comprehension are described, and common features of the cognitive processes underlying text comprehension are highlighted to provide the basis for further research. ...
Added: September 28, 2026
Sherbakova O., Обольская В. А., Вопросы психолингвистики 2023 Т. 55 № 1 С. 112–127
The article describes the second part of the research aimed at developing the framework for studying the comprehension of implicit meanings of verbal texts. The uniqueness of the proposed framework stems from the combination of psychological and linguistic methodologies. We outline the current approaches to the phenomenon of text comprehension along with its classifications available ...
Added: September 28, 2026
Sherstinova T., Kirina M., Herald of the Russian Academy of Sciences 2026 Vol. 96 P. 40–49
The paper presents the results of a series of studies on the emotional dimension of the Russian short story of the first third of the 20th century. The study is based on a sample of 210 texts from the Corpus of the Russian Short Story of 1900–1930. Two fundamentally different approaches are compared: automatic detection ...
Added: September 11, 2026
Долгополова Ю. Н., Kotov A. A., Юсупова М. А. et al., М.: НИУ ВШЭ, Институт образования, 2026.
More than half of Russian adolescents use generative neural networks daily, and this figure continues to grow. Under these conditions, AI is becoming not just a tool but a cognitive partner for students, which requires a rethinking of existing educational practices. This systematic review synthesizes empirical data from studies conducted between 2014 and 2026 on ...
Added: August 21, 2026
Trofimova E., Shamina Z., Selifanova M. et al., , in: Proceedings of the Generative Code Intelligence Workshop (GeCoIn 2026), co-located with the 35th International Joint Conference on Artificial Intelligence (IJCAI-ECAI 2026)Vol. 4238.: CEUR-WS.org, 2026.
We introduce ML2B, the first benchmark for evaluating cross-lingual task comprehension in end-to-end ML pipeline generation by large language models. Despite growing global AI adoption, no systematic evaluation exists for ML pipeline generation beyond English task descriptions. ML2B addresses this gap with 35 Kaggle competitions spanning tabular, text, and image domains, translated into 14 languages ...
Added: August 20, 2026
Surkov A., Ignatenko V., Koltsov S., Computers, Materials and Continua 2026 Vol. 88 No. 3 Article 74
Large language models have recently demonstrated promising capabilities in mathematical reasoning; however, their performance on tasks requiring strict symbolic manipulation, such as solving differential equations, remains limited, especially for compact models. In this work, we investigate whether activation steering combined with reinforcement learning can improve the quality of solutions generated by pretrained language models without ...
Added: July 8, 2026
Stognieva O., Чеснокова Н. Е., Отечественная и зарубежная педагогика 2026 Т. 1 № 3 (115) С. 123–131
Integration of generative artificial intelligence tools into educational practice highlights the need for pedagogically grounded approaches to their use in the creation of educational video content, which is increasingly applied in language and professionally oriented instruction.
The purpose of this article is to conduct a comparative analysis of educational video content created using generative AI tools ...
Added: June 20, 2026
Parkina U., Rakhuba M., , in: 39th Conference on Neural Information Processing Systems (NeurIPS 2025).: NeurIPS, 2025. P. 71014–71041.
Recent studies suggest that context-aware low-rank approximation is a useful tool for compression and fine-tuning of modern large-scale neural networks. In this type of approximation, a norm is weighted by a matrix of input activations, significantly improving metrics over the unweighted case. Nevertheless, existing methods for neural networks suffer from numerical instabilities due to their ...
Added: April 29, 2026
Fedorov A., Карпов С. В., Библиосфера 2025 № 4 С. 106–115
Artificial intelligence (AI) transforms information search by providing librarians with new effective tools to assist users. It changes the ways of information retrieval, where librarians play an important role in helping their users. The purpose of this article is to examine the perception of artificial intelligence by domestic librarians. Based on the analysis of publications, ...
Added: March 2, 2026
Timofeyeva O. A., Kushkin N. A., Мониторинг общественного мнения: Экономические и социальные перемены 2024 № 5 С. 60–84
Artificial intelligence (AI) has long been a part of our lives, however, mass access to ChatGPT and Midjourney in 2022 has taken the discussion about it to a new level. This article examines the perception of AI as a technology in general, as well as its possible roles and practices of use, by students, teachers, ...
Added: February 27, 2026
Karavaeva E., Vasilevsky V., Ланин Г. М. et al., Труды Института системного программирования РАН 2025 Т. 37 № 4 С. 175–190
The ongoing digitalization of education requires new ways of presenting information and attention retention mechanisms. The aim of the presented work is to propose a solution for implementing a large language model, which will interactively generate prompts of different types, within an e-learning course on programming. The main approaches are the analysis of existing relatively ...
Added: December 25, 2025
Fedorov A., Современная библиотека 2025 № 8 С. 20–23
Представьте, что вы заходите в библиотеку и надеваете очки виртуальной реальности, которые переносят вас в Древний Китай, а ИИ-гид рассказывает о бамбуковых дощечках, шелковых свитках и о том, как формировались первые императорские книжные собрания. Или вы оказываетесь в просторном зале виртуального музея, где оживают картины великих художников. Вы приближаете детали на полотнах, сравниваете стили, а ...
Added: November 20, 2025