Iskanje po repozitoriju

Opis: This paper presents a baseline study on automatic Uzbek text classification. Uzbek is a morphologically rich and low-resource language, which makes reliable preprocessing and evaluation challenging. The approach integrates Term Frequency–Inverse Document Frequency (TF–IDF) representation with three conventional methods: linear regression (LR), k-Nearest Neighbors (k-NN), and cosine similarity (CS, implemented as a 1-NN retrieval model). The objective is to categorize school learning materials by grade level (grades 5–11) to support improved alignment between curricular texts and students’ intellectual development. A balanced dataset of Uzbek school textbooks across different subjects was constructed, preprocessed with standard NLP tools, and converted into TF–IDF vectors. Experimental results on the internal test set of 70 files show that LR achieved 92.9% accuracy (precision = 0.94, recall = 0.93, F1 = 0.93), while CS performed comparably with 91.4% accuracy (precision = 0.92, recall = 0.91, F1 = 0.92). In contrast, k-NN obtained only 28.6% accuracy, confirming its weakness in high-dimensional sparse feature spaces. External evaluation on seven Uzbek literary works further demonstrated that LR and CS yielded consistent and interpretable grade-level mappings, whereas k-NN results were unstable. Overall, the findings establish reliable baselines for Uzbek educational text classification and highlight the potential of extending beyond lexical overlap toward semantically richer models in future work.
Ključne besede: Uzbek language, text classification, low-resource languages, TF-IDF, cosine similarity, linear regression, k-Nearest Neighbors
Objavljeno v RUP: 17.10.2025; Ogledov: 507; Prenosov: 4
Celotno besedilo (286,87 KB)
Gradivo ima več datotek! Več...

Opis: The aim was to evaluate the effectiveness of semen cryopreservation and oocyte vitrification in roe deer as a potential method of gamete preservation for endangered deer species. Sperm were isolated from the cauda epididymis of fourteen bucks (n = 14). The motility measure (CASA) and morphology of fresh semen (FS) and frozen–thawed semen (TS) were compared. A hyaluronic binding assay was used to distinguish between mature FS spermatozoa expressing hyaluronan receptors and immature FS lacking these receptors, and the mitochondrial membrane potential (MMP) in TS was determined (flow cytometry). A Sperm–Hyaluronan Binding Assay (HBA) showed a viability rate of 61.9% in FS and 78.2% in TS. Oocytes received from eight does (n = 8) underwent a viability test and vitrification, and fresh oocytes from the other eight does (n = 8) were fertilized with TS and embryos were cultured until the blastocyst stage. The number of isolated oocytes, cumulus–oocyte complexes (COCs), cleaved embryos, and expanded blastocysts was evaluated. Higher percentages of morphological factors (acrosome, head, midpiece, and tail shape) were observed in FS compared to TS, whereas the motility and progressive movement were greater in TS (p ≤ 0.001). The viability was 50.5% and MMP was 40.6% in TS. A total of 311 oocytes were collected and from 150 COCs and 125 blastocysts developed. The viability of thawed oocytes after vitrification was 81%. The viability of vitrified oocytes and cryopreserved sperm confirmed the effectiveness of freezing protocols and highlights the potential for their implementation in other deer species.
Ključne besede: roe deer, semen, oocytes, fertilization
Objavljeno v RUP: 13.10.2025; Ogledov: 461; Prenosov: 14
Celotno besedilo (1,10 MB)
Gradivo ima več datotek! Več...

Opis: Brain MRI segmentation plays a crucial role in neuroimaging studies and clinical trials by enabling the precise localization and quantification of brain tissues and structures. The advent of deep learning has transformed the field, offering accurate and fast tools for MRI segmentation. Nevertheless, several challenges limit the widespread applicability of these methods in practice. In this systematic review, we provide a comprehensive analysis of developments in deep learning-based segmentation of brain MRI in adults, segmenting the brain into tissues, structures, and regions of interest. We explore the key model factors influencing segmentation performance, including architectural design, choice of input size and model dimensionality, and generalization strategies. Furthermore, we address validation practices, which are particularly important given the scarcity of manual annotations, and identify the limitations of current methodologies. We present an extensive compilation of existing segmentation works and highlight the emerging trends and key results. Finally, we discuss the challenges and potential future directions in the field.
Ključne besede: magnetic resonance imaging, brain, image segmentation, deep learning
Objavljeno v RUP: 10.10.2025; Ogledov: 792; Prenosov: 12
Celotno besedilo (956,69 KB)
Gradivo ima več datotek! Več...

Opis: This paper describes an open-source computational tool developed for the modeling and simulation of fed-batch bioreactors, particularly for processes employing Chinese Hamster Ovary (CHO) cells, which are integral to biopharmaceutical manufacturing. The software provides a platform for researchers and industry professionals to simulate bioreactor dynamics and investigate the impact of various operational parameters, such as nutrient supply rates, oxygen concentrations, and temperature, prior to physical experimentation. The tool enables users to generate predictions of critical variables including cell density, nutrient consumption, and product concentration profiles over time. These predictions are derived from a mathematical framework based on a system of ordinary differential equations solved using the Runge–Kutta method. A notable capability of the software is the import of experimental data and the application of the Nelder–Mead algorithm for parameter optimization, allowing for the calibration of the model against empirical findings, thereby enhancing its predictive accuracy. The software supports in silico experimentation, which can contribute to reducing the time, cost, and resources associated with optimizing bioreactor configurations and scaling up production processes. By providing a refined and adaptable framework, this instrument assists in improving the understanding of bioreactor dynamics, optimizing biopharmaceutical production methodologies, and correlating theoretical models with practical bioreactor operations. The software is available as an open-source project to promote its adoption and continued development within the scientific community.
Ključne besede: bioreaktor, mathematical modeling, CHO
Objavljeno v RUP: 29.09.2025; Ogledov: 707; Prenosov: 5
Celotno besedilo (939,94 KB)
Gradivo ima več datotek! Več...

Opis: In deep learning-based side-channel analysis (DL-SCA), there may be a proliferation of model parameters as the number of trace power points increases, especially in the case of raw power traces. Determining how to design a lightweight deep learning model that can handle a trace with more power points and has fewer parameters and lower time costs for profiled SCAs appears to be a challenge. In this article, a DL-SCA model is proposed by introducing a non-trained DL technique called random convolutional kernels, which allows us to extract the features of leakage like using a transformer model. The model is then processed by a classifier with an attention mechanism, which finally outputs the probability vector for the candidate keys. Moreover, we analyze the performance and complexity of the random kernels and discuss how they work in theory. On several public AES datasets, the experimental results show that the number of required profiling traces and trainable parameters reduce, respectively, by over 70% and 94% compared with state-of-the-art works, while ensuring that the number of power traces required to recover the real key is acceptable. Importantly, differing from previous SCA models, our architecture eliminates the dependency between the feature length of power traces and the number of trainable parameters, which allows for the architecture to be applied to the case of raw power traces.
Ključne besede: side-channel analysis, deep learning, convolution neural networks, random convolution kernel
Objavljeno v RUP: 26.09.2025; Ogledov: 1680; Prenosov: 7
Celotno besedilo (1,75 MB)
Gradivo ima več datotek! Več...