This article discusses the practical issues of compiling controlled multilingual thesauri for the purposes of industry-specific translation (IST). In the multilingual, transnational and globally connected Kazakhstan, IST is a much-needed translation service. IST is an interdisciplinary field between terminology, computational linguistics, translation theory and practice. Most of the professional guides, dictionaries and glossaries are systemized in alphabetical order and contain multiple variants for the terms searched. Therefore, there is an urgent need to create a systemized controlled multilingual thesaurus of industry-specific Kazakh, English and Russian terms in order to provide multilingual users with an interoperable and relevant term base. Controlled multilingual thesauri for industry-specific terms are the most effective tools for describing individual subject areas. They are designed to promote communication and interaction among professionals, translators and all Automated Information System users of specific fields irrespective of their location and health conditions. Unlike traditional dictionaries, controlled thesauri allow users to identify the meaning with the help of definitions and translations, relations of terms with other concepts, and broader and narrower terms. The purpose of this research is to unify and systematize industry-specific terms in Kazakh, to provide Russian and English equivalents, and to classify the terms into essential rubrics and subjects. Based on the Zthes data scheme to create a controlled multilingual thesaurus of industryspecific terms, the major rubrics have been formulated, and about 10,000 Kazakh mining and metal terms approved by the Terminological Committee of Kazakhstan have been structured.
This paper examines grammatical transformations (translation techniques), in rendering the works of Akhmet Baitursynuly into Russian, using a corpus-based approach. The primary analysis is based on the Kazakh-Russian parallel corpus of Baitursynuly’s texts, created by the authors. This corpus serves as a foundational tool for identifying linguistic patterns, translation techniques, and stylistic nuances across the languages. The study focuses on selected texts by Baitursynuly in both Kazakh and Russian. Focusing on selected original texts and their translations, the identified key grammatical transformations include syntactic assimilation, sentence splitting and merging, and grammatical substitutions. These transformations are analyzed to uncover their role in adapting linguistic structures and preserving the stylistic essence of the source material. The study highlights the utility of parallel corpora in translation studies, demonstrating their effectiveness in analyzing linguistic and stylistic nuances. A frequency analysis of specific translation challenges, such as the causative conjunctions sebebі and üjtkenі (both meaning because) in Kazakh, reveals variations in their Russian equivalents influenced by linguistic constraints and Baitursynuly’s stylistic preferences. Corpus-based research methods enable sampling, comparative and contrastive text analysis, and detailed examination of grammatical transformations. The parallel corpus facilitates the alignment of source and target texts, offering quantitative and qualitative insights into the translation process. This approach also provides an empirical framework for examining recurring translation patterns and assessing the influence of language typology on grammatical adaptation. The study utilized exhaustive sampling of linguistic material from original texts and translations, comparative and contrastive text analysis, analysis of grammatical transformations, and text alignment.
This study focuses on the classification of Kazakh text data using deep learning models. Text classification constitutes a foundational component within the domain of natural language processing (NLP), frequently serving as a preliminary step in numerous other text-processing applications. The research considered the primary task of text data classification, i.e. sentiment analysis. The study proposed the use of deep learning models for the automated classification of text data into positive and negative opinions. Since the Kazakh is a low-resource language, it is imperative to create specialised datasets, perform preprocessing of text, and carry out annotation. The study utilised the KazSAnDRA dataset, in conjunction with a dataset prepared by the authors. A range of deep learning models, including CNN, LSTM, GRU, and DistilBERT, were utilised in the experimental evaluations. Among the models examined, the DistilBERT and LSTM models demonstrated the highest levels of accuracy and were identified as a promising approach for the automatic classification of Kazakh texts. The findings of this study have the potential to provide a theoretical and practical foundation for future research endeavours focused on the analysis of texts in the Kazakh language.
This paper presents an adaptive image steganography technique that integrates linear hash functions with chessboard-partitioned image blocks to achieve high-capacity data embedding while preserving visual quality. The proposed method operates on 8×8 pixel blocks, which are classified based on local variance into three categories: smooth, moderately complex, and highly textured. Depending on the category, the algorithm adaptively determines the number of least significant bits (LSBs) used for embedding, allowing for a balance between payload and imperceptibility. A key innovation of the approach is the use of staggered partitioning within each block, where one subset of pixels is used for embedding, and the other remains unchanged to enable blind estimation of variance during extraction. This enables the decoder to reconstruct the embedding conditions without requiring access to the original image or auxiliary metadata. The embedding process utilizes a linear hash function based on primitive polynomial division in the Galois Field GF(2). This allows an m-bit message to be embedded into an $\hat D = \left\{ {{s_1}, \ldots {s_N}} \right\}$ bit sequence with only a single-bit modification, ensuring minimal statistical disruption to the cover image. Experimental validation was conducted using the BOSSBase v1.01 dataset, where the method achieved embedding rates ranging from 0.3 to 0.9 bits per pixel (bpp), with Peak Signal-to-Noise Ratio (PSNR) values consistently exceeding 44 dB. In addition to high fidelity and capacity, the algorithm is computationally efficient, relying on simple binary operations such as XOR and polynomial transformations in GF(2), which makes it suitable for implementation in low-resource environments such as embedded and mobile systems.
About Me
English-Kazakh translator, PhD in Translation Studies, Applied linguist
English-Kazakh translator, PhD in Translation Studies, Applied linguist