Abstract
Knowledge models often use hierarchical structures, which help break down complex data into manageable components. This enables better understanding and aids in reasoning and decision-making. Hierarchical structures are effective in organizing, managing, and processing complex information. Traditional Self-Organizing Maps are typically flat, two-dimensional grids for visualizing and grouping data. They can be shaped into hierarchical structures, offering benefits such as improved data representation, scalability, enhanced grouping and visualization, and hierarchical feature extraction while preserving data topology. This paper introduces a self-organizing hierarchical map with an appropriate topology and a suitable learning mechanism for retaining information in an organized way. In this conceptual model, information is selectively absorbed in each layer. These characteristics make the Hierarchical Self-organising Maps a powerful non-linear classifier. Simulations are conducted to test and evaluate the performance of this neural structure as a classifier.

Abstract
Solar wind forecasting is a core component of Space Weather, a field that has been the target of many novel machine-learning approaches. The continuous monitoring of the Sun has provided an ever-growing ensemble of observations, facilitating the development of forecasting models that predict solar wind properties on Earth and other celestial objects within the solar system. This enables us to prepare for and mitigate the effects of solar wind-related events on Earth and space. The performance of some simulation-based solar wind models depends heavily on the quality of the initial guesses used as initial conditions. This work focuses on improving the accuracy of these initial conditions by employing a Recurrent Neural Network model. The study's findings confirmed that Recurrent Neural Networks can generate better initial guesses for the simulations, resulting in faster and more stable simulations. In our experiments, when we used predicted initial conditions, simulations ran an average of 1.08 times faster, with a statistically significant improvement and reduced amplitude transients. These results suggest that the improved initial conditions enhance the numerical robustness of the model and enable a more moderate integration time step. Despite the modest improvement in simulation convergence time, the Recurrent Neural Networks model's reusability without retraining remains valuable. With simulations lasting up to 12 h, an 8% gain equals one hour saved per simulation. Moreover, the generated profiles closely match the simulator's, making them suitable for applications with less demanding physical accuracy.

Abstract
This study addresses the sonar perceptual ambiguity problem inherent in underwater sonar-based SLAM systems, resulting from the sonar's wide beam aperture. We propose a SLAM algorithm that integrates a free-space mapping approach with a particle filter, leveraging polar-based acoustic images collected from an MSIS. Our method utilizes grid maps to compile information on empty regions, aiming at improving mapping accuracy. Free-space representations are further leveraged for particle filter weight update via scan matching. The centre-of-mass map cell representation is exploited for efficient weight update using simple matrix operations. Illustrative experimental results are provided, based on real data collected from a testing pool environment.

Abstract
One of the key tasks in ensuring national security is the ability of the state and society to recognise and effectively assess the conditions for disasters, and to prevent them from threatening the sustainable development of the country. The Kyrgyz Republic is highly vulnerable to the influence of climate change, which in turn affects the frequency and intensity of disasters. The Kyrgyz Republic is exposed to almost all types of geological and man-made hazards, including earthquakes, landslides, debris flows, flash floods, outbursts of mountain lakes, dam failures, avalanches, droughts, extreme temperature, epidemics and releases of hazardous substances. Analysis of information on existing risks and their control systems used to reduce their negative impact makes it possible to assess the degree of probability, the expected consequences of threats, determine the degree of risk, the adaptive potential of communities and select appropriate protective measures. Therefore, this study is conducted to assess the hazard, vulnerability and exposure of Suzak district (Jalal-Abad oblast) in order to quantify the risk of the study area using multi-parameter holistic assessment with field collecting of primary data and utilizing Index-based Risk Assessment approach based on applying INFORM Risk model. Collected data was used to downscale subnational INFORM Risk model for municipal and district level using a multi-layered structure. A risk score is calculated by combining 72 indicators that measure three main dimensions: hazard & exposure, vulnerability, and lack of coping capacity. These findings provide an opportunity to develop a more effective disaster risk management at the local and national levels, by prioritizing relevant actions and investments for municipalities – districts which are demonstrated relatively highest risk scores. Also, the possibility of applying localized risk assessment procedures provides an opportunity to obtain more accurate sub-national (district/oblast based) and national levels with effective assessing dynamics of risk.

Abstract
Geophysical studies on the lunar surface have, in the past, used various methods that contribute not only towards the knowledge of the lunar subsurface but also contribute towards the design of future lunar missions, namely those that will, in the near future, take humans to the Moon’s surface. This work analyzes a specific set of ground penetrating radar (GPR) data, collected during the Chang’E-4 mission of the Chinese Space Agency, using theYutu-2 rover within the von Kármán crater, on the far-side of the Moon. From this dataset two electrical parameters were estimated. The approach uses the backscatter of the electromagnetic wavefield in order to obtain estimates of the real component of the complex relative permittivity as well as the electrical resistivity. © 2024 International Multidisciplinary Scientific Geoconference. All rights reserved.

Abstract
Raw materials are essential for Europe’s industrial base, as they are used to produce vital goods and technologies. The European Comission’s assessment identifies lithium and tin as critical and strategic raw materials due to rising demand. A PhD thesis aims to create a 3D geological model of the Argemela District in Central Portugal to provide vital information about the genesis of the hydrothermal tin and lithium mineralizations founded in the region. The study places emphasis on the very-low-frequency (VLF) electromagnetic method as a tool to provide information about the mineralization, lithologic contacts, and structural features that can be related to the mineralizations. Argemela district has two main areas, the Argemela Tin and Lithium Mine and the Argemela Hill Top. VLF data was collected and analyzed, showing that low resistivity may be associated with mineralization in the Argemela Mine, while high resistivity may be linked to the Argemela microgranite in the Argemela Hill Top. This geophysical method is effective in non-invasively mapping subsurface features, assisting in the development of a comprehensive 3D geological model and enhancing resource evaluation.