Graph approaches are becoming increasingly popular in various areas of life and provide a powerful tool for displaying complex relationships between different modalities of data. This study considers a new graph document-oriented structure for processing multimodal medical data in order to improve disease prediction and analysis of patient treatment outcomes. The study emphasizes the importance of using graph models in health care and demonstrates their advantages for processing large and heterogeneous data sets, which allows obtaining more accurate and prompt medical conclusions. In particular, medical images, electronic health records (EHR), records from wearable devices, environmental records, psycho-emotional state and others are taken into account. The proposed data processing pipeline includes several stages: preprocessing, creation of a base graph, minimization of the number of edges, recalculation of edge weights. In the future, it is advisable to apply graph neural network (GNN) approaches, in particular Graph Convolutional Networks (GCN) and Graph Attention Networks (GAT), for better integration of internal data, which will allow for a more comprehensive view of patient data.
Melanoma, an aggressive malignancy of melanocytes, is one of the deadliest skin cancers, marked by high mortality rates, particularly when diagnosed late. This study presents the development of an autonomous diagnostic system designed to detect melanoma using embedded AI technologies, specifically convolutional neural networks and advanced image processing methods. The system aims to enhance diagnostic accuracy, shorten waiting times for diagnosis, and provide a non-invasive, accessible solution for early melanoma detection. We leverage deep learning models trained on a diverse dataset of dermoscopic images, combined with innovative pre-processing and segmentation techniques, to achieve high-performance melanoma classification. The results demonstrate the potential of this web-based system to serve as an effective decision support tool for clinicians, ultimately improving patient outcomes through early intervention.
Deep learning models, such as Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), have achieved state-of-the-art performance in the classification of brain lesions on MRI images. However, the complexity of this type of images requires CNNs to use deeper architectures with more parameters to effectively capture their high-dimensional features and subtle variations. On one side, ViTs provide a different approach to tackle this challenge, but they require larger datasets and more computational costs. On the other side, ensemble deep learning techniques, such as Bagging, Stacking, and Boosting, can help mitigate these limitations by combining multiple CNN models. This study explores these methods and compares them to evaluate their accuracy and efficiency using three approaches: CNN-based transfer learning, ViT-based transfer learning, and ensemble deep learning techniques, such as Bagging, Stacking, and Boosting based on XGBoost, AdaBoost methods. The experiments, carried out on four MRI image datasets with different levels of complexity and types of brain lesions, show that the combination of CNNs with ensemble techniques offers very competitive performances to those of individual CNNs and ViTs, with interesting improvements compared to already existing approaches.
The imputation process for genetic data is cost and time-intensive, primarily due to the high complexity of the methods involved, and the substantial volume of data processed. A thorough performance evaluation of the imputation algorithms such as Beagle, AlphaPlantImpute, LinkImputeR, MACH and others shows that while some algorithms are highly accurate, they are often computationally expensive. Being widely used, they have multiple input parameters which impact the quality and accuracy of the imputation. Traditional machine learning techniques for parameter optimization like grid search and randomized search become inefficient in high-dimensional parameter spaces, leading to prohibitive computational costs, especially in large-scale applications. Our study proposes the cloud-based approach for input parameters optimization by using Bayesian optimization with consecutive Domain Reduction Transformer (DRT). Described algorithm and developed library allow users to find the optimal input parameters for the data imputation in a more flexible way.
The article addresses the pressing and significant issues of managing the process of transport evacuation for those affected by military actions, as well as managing emergency medical assistance for the wounded. Amidst the ongoing active warfare in Ukraine, there is a growing need to integrate the efforts of all stakeholders-on the one hand, those overseeing these operations at all levels of the hierarchy, and on the other hand, those directly executing them in the field where emergency aid is provided. According to the authors, effective management of these processes is feasible through a model that not only swiftly responds to constant changes but also facilitates efficient coordination among stakeholders. Such a model is the Architecture-Oriented Agent-Based Model (AOAM), which meets all the requirements of crisis management. The research includes a parametric analysis of existing models to identify the optimal one, considering a combination of factors, key characteristics, and limitations. The authors also explore the challenges of developing and implementing AOAM, given the scarcity of financial and technical resources, and propose alternative approaches for project execution. The article examines the potential for expanding the functionalities of the proposed AOAM through its integration with GIS technologies. Furthermore, the research investigates the prospective applications of this integrated model in remediation operations and its use in peacetime.
