The article is devoted to the creation of an effective solution for user segmentation. The article presents an analysis of existing user segmentation services, an analysis of approaches to user segmentation (ABCDx segmentation, demographic segmentation, segmentation based on a user journey map), an analysis of clustering algorithms (K-means, Mini-Batch K-means, DBSCAN, Agglomerative Clustering, Spectral Clustering). The study of these areas is aimed at creating a “flexible” segmentation solution that adapts to each user sample. Dispersion analysis (ANOVA test), analysis of clustering metrics is also used to assess the quality of user segmentation. With the help of these areas, an effective solution for user segmentation has been developed using advanced analytics and machine learning technology.

This paper presents a neuro-fuzzy algorithm for high-speed grayscale image segmentation based on a modified defuzzification method using triangular membership functions. The aim of the study is to analyze the effect of simplifying the defuzzification formula on the accuracy and contrast of object selection. The proposed approach includes adaptive learning of the weight coefficient, which allows dynamically adjusting the defuzzification process depending on the target values. The paper compares the basic method of averaging membership values and a modified version taking into account nonlinear weights. Experiments conducted on 1024x720 images demonstrate that the developed algorithm provides high segmentation accuracy and improved object contrast with minimal computational costs. The results confirm the superiority of the proposed method over traditional approaches, emphasizing the prospects for applying artificial intelligence in computer vision problems.

A comparative analysis of image segmentation methods for fire detection was conducted using thresholding in the HSV color space and the U-Net neural network. The study aimed to evaluate the efficiency of these approaches in terms of execution time and fire detection accuracy based on RMSE, IoU, Dice, and MAPE metrics. Experiments were performed on four different fire images with manually prepared ground truth fire masks. The results showed that the HSV method offers high processing speed (0.0010–0.0020 s) but tends to detect not only fire but also smoke, reducing its accuracy (IoU 0.0863–0.3357, Dice 0.1588–0.5026). The U-Net neural network demonstrates higher fire segmentation accuracy (IoU up to 0.6015, Dice up to 0.7512) due to selective flame detection but requires significantly more time (1.2477–1.3733 s) and may underestimate the total fire area (MAPE up to 78.5840%). Visual assessment confirmed differences in methods' behavior: HSV captures smoke as part of the target area, while U-Net focuses exclusively on fire. The choice between methods depends on task priorities: speed or accuracy. Future research directions were proposed, including U-Net optimization and the development of hybrid approaches.

Handwritten archival documents form a fundamental part of humanity's cultural heritage. However, their analysis remains a labor-intensive task for professional researchers, such as historians, philologists, and linguists. Unlike commercial OCR applications, working with historical manuscripts requires a fundamentally different approach due to the extreme diversity of handwriting, the presence of corrections, and material degradation.

This paper proposes a method for searching within handwritten texts based on stroke segmentation. Instead of performing full text recognition, which is often unattainable for historical documents, this method allows for efficiently answering researcher search queries. The key idea involves decomposing the text into elementary strokes, forming semantic vector representations using contrastive learning, followed by clustering and classification to create an adaptive handwriting dictionary.

It is experimentally shown that search by comparing tuples of reduced sequences of the most informative strokes using the Levenshtein distance provides sufficient quality for the task at hand. The method demonstrates resilience to individual handwriting characteristics and writing variations, which is particularly important for working with authors' archives and historical documents.

The proposed approach opens up new possibilities for accelerating scientific research in the humanities, reducing the time required to find relevant information from weeks to minutes, thereby qualitatively transforming research capabilities when working with large archives of handwritten documents.

Steel is one of the most important bulk materials these days. It is used almost everywhere - from medicine to industry. Detecting this material's defects is one of the most challenging problems for industries worldwide. This process is also manual and time-consuming. Through this study we tried to automate this process. A convolutional neural network model UNet was used for this task for more accurate segmentation with less training image data set for our model. The essence of this NN (neural network) is in step-by-step convolution of every image (encoding) and then stretching them to initial resolution, consequently getting a mask of an image with various classes on it. The foremost modification is changing an input image's size to 128x800 px resolution (original images in dataset are 256x1600 px) because of GPU memory size's limitation. Secondly, we used ResNet34 CNN (convolutional neural network) as encoder, which was pre-trained on ImageNet1000 dataset with modified output layer - it shows 4 layers instead of 34. After running tests of this model, we obtained 92.7% accuracy using images of hot-rolled steel sheets.

The complexity of the problem of extracting argumentative structures is associated with such problems as selecting argumentative segments, predicting long-range connections between non-contact segments, and training on data labeled with a low degree of inter-annotator consistency. In this paper, we consider an approach to extracting argumentative relations from fairly large texts related to scientific communication. A comparative analysis was performed of fine-tuning methods using a pre-trained Longformer-type language model that takes into account long contexts and two methods that take into account annotator discrepancies in argument labeling by using the so-called soft labels obtained by uniformly smoothing labels and averaging expert assessments. The experiments were conducted on four datasets containing positive and negative examples of statement pairs (premise, conclusion) and differing in segmentation methods and average text size. The best results were obtained using the model with averaging expert assessments. At the same time, it is noted that the model using smoothed labels also increases the accuracy of classifiers, but worsens the recall.

The study examines the principle of creating virtual exhibitions as a means of integration into the Common Digital Space of Scientific Knowledge (CDSSK), information systems in the field of science and culture, with the aim of promoting science, ensuring access to information in various scientific fields, and drawing attention to current issues and achievements in the scientific sphere. The main methods of creating virtual exhibitions are formulated, including content selection and segmentation into main sections. In addition, a classification of virtual exhibitions into autonomous, remote, and combined is proposed. Special attention is paid to the methodology of creating virtual exhibition at the Moscow Center of the Russian Academy of Sciences. Using the example of an interdepartmental combined virtual exhibition, a detailed description of the "Mrs. Penicillin" exhibition dedicated to the creator of penicillin, Z.V. Ermolyeva, is provided.