Search

In existing models and embedding methods of networked systems, node features describing their qualities are usually overlooked in favor of focusing solely on node connectivity. This study introduces $FiD$-Mercator, a model-based ultra-low dimensional reduction technique that integrates node features with network structure to create $D$-dimensional maps of complex networks in a hyperbolic space. This embedding method efficiently uses features as an initial condition, guiding the search of nodes' coordinates towards an optimal solution. The research reveals that downstream task performance improves with the correlation between network connectivity and features, emphasizing the importance of such correlation for enhancing the description and predictability of real networks. Simultaneously, hyperbolic embedding's ability to reproduce local network properties remains unaffected by the inclusion of features. The findings highlight the necessity for developing novel network embedding techniques capable of exploiting such correlations to optimize both network structure and feature association jointly in the future.

Graph Neural Networks (GNNs) have excelled in predicting graph properties in various applications ranging from identifying trends in social networks to drug discovery and malware detection. With the abundance of new architectures and increased complexity, GNNs are becoming highly specialized when tested on a few well-known datasets. However, how the performance of GNNs depends on the topological and features properties of graphs is still an open question. In this work, we introduce a comprehensive benchmarking framework for graph machine learning, focusing on the performance of GNNs across varied network structures. Utilizing the geometric soft configuration model in hyperbolic space, we generate synthetic networks with realistic topological properties and node feature vectors. This approach enables us to assess the impact of network properties, such as topology-feature correlation, degree distributions, local density of triangles (or clustering), and homophily, on the effectiveness of different GNN architectures. Our results highlight the dependency of model performance on the interplay between network structure and node features, providing insights for model selection in various scenarios. This study contributes to the field by offering a versatile tool for evaluating GNNs, thereby assisting in developing and selecting suitable models based on specific data characteristics.

U istraživanju se ispitivao seizmički sudar dviju susjednih zgrada uzimajući u obzir interakciju tla i konstrukcije (engl. soil-structure interaction - SSI). Provedena je sveobuhvatna parametarska analiza zgrada različitih visina kako bi se ispitalo ponašanje uzrokovano sudarima, uzimajući u obzir učinak tla. Provedena je valićna transformacija (engl. wavelet transform) kako bi se dobio uvid u razlike u učestalosti udarnih sila između susjednih konstrukcija s nepomičnom i s fleksibilnom bazom. Linearni viskozni prigušivači (LVD) između susjednih katova upotrijebljeni su kao mjere zaštite od sudara. Potreban dodatni omjer prigušenja LVD-ova određen je optimizacijskom analizom pri različitim vrstama tla kako bi se potvrdio učinak SSI-ja na konstrucijsko prigušenje. Usporedni rezultati s SSI-jem i bez njega pokazali su da uključivanje SSI-ja pogoršava odzive uzrokovane sudarima tijekom potresa., This study investigated the seismic pounding of two adjacent buildings considering soil–structure interaction (SSI). A comprehensive parametric study of buildings with different heights was performed to reveal the pounding-involved behaviour considering the soil effect. Wavelet transform has been conducted to gain insight into the differences in the frequency contents of the impact forces between fixed- and flexible-base adjacent structures. Linear viscous dampers (LVDs) between adjacent floors were used as pounding protection measures. The required supplemental damping ratio of the LVDs was determined through optimisation analysis under different soil types to verify the effect of the SSI on structural damping. Comparative results with and without SSI showed that incorporating the SSI worsens the pounding-involved responses during earthquakes.

Every Marine is a rifleman. However, the Marine Corps has failed to leverage technology to improve the art and science of marksmanship training, which has gone unchanged for almost 100 years. Marines score paper targets with pen and paper, wasting valuable time. The data collected is insufficient to precisely measure shooter abilities or evaluate analytically. The result is a lack of infrastructure and programs to leverage data analytics, machine learning, and artificial intelligence to aid in coaching Marines. This thesis expands on a previous Naval Postgraduate School thesis by continuing the development of tools to exploit digital marksmanship data to gain meaningful insights. With these tools, Marines would have unparalleled access to their historical shooting data and coaching feedback. Quantifiable measures will be used to provide coaching recommendations to improve rifle employment by providing more focused and effective training. We identify a more accurate method for calculating a probability of hit from observed data, which will allow the Marine Corps to measure lethality more accurately. We also offer recommendations for resource allocation to training facilities for marksmanship modernization and for programmatic requirements for automated data collection, storage, and use in analysis. By harnessing analytics and artificial intelligence, the Marines will more efficiently and effectively train Marines to be riflemen and enhance the lethality of the USMC., Major, United States Marine Corps, Approved for public release. distribution is unlimited

Using a set of over 70.000 records from PLOS One journal consisting of 37 lexical, sentiment and bibliographic variables we perform analysis backed with machine learning methods to predict the class of popularity of scientific papers defined by the number of times they have been viewed. Our study shows correlations among the features and recovers a threshold for the number of views that results in the best prediction results in terms of Matthew's correlation coefficient. Moreover, by creating a variable importance plot for random forest classifier, we are able to reduce the number of features while keeping similar predictability and determine crucial factors responsible for the popularity., Comment: 13 pages, 6 figures

Grandstands are structures which are regularly subjected to dynamic loads generated by crowd motions. It is a dangerous situation when spectators induce rhythmic jumping, dancing, stamping, etc. If the synchronized movement of spectators excites a natural frequency of the structure, resonant response might occur. To avoid such situations, temporary steel grandstands are commonly strengthened using additional elements that create bracing system which is selected depending on the size of the structure, type of the event, acting load, etc. It was proved that not only the use but also the arrangement of such structural members is crucial for the dynamic structural resistance. The aim of the present study is to determine the most effective arrangement of bracing system for a typical example of the temporary steel grandstand which is exposed to dynamic load induced by spectators. Three different arrangements of bracing systems have been analysed using five criteria recommended in the literature. The results of the study clearly show that the dynamic parameters of the grandstand are substantially different for various types of bracing systems. The largest improvement in the structural behaviour has been obtained for the grandstand equipped with the bracing system satisfying all proposed criteria. The peak accelerations for this case have been found to be nearly twice as low as for structures with other bracing arrangements. The application of such a system for the grandstand which is exposed to human-induced vibrations allows for safe use of the structure as well as improves comfort of spectators.

The paper deals with the stability of earth dams subjected to seismic and paraseismic excitations occurring in the close distance to the dam. The Newmark’s stability approach, classified as a simplified dynamic method, has been extended for two directional dynamic loading case and the evolution of the friction coefficient. The paper is focused on the experimental verification of the proposed method for harmonic and cyclic excitations generated by a simple oscillating device as well as by one-directional shaking table. Main assumptions of the proposed method have been proved and the effect of the vertical acceleration on the stability of the dam has been demonstrated. In the experiments conducted under cyclic loading, the impact of the vertical acceleration on the permanent displacement of the block was equal to 57 %. However, in general, the exact magnitude of this impact depends on many factors, such as the vertical to horizontal peak ground acceleration ratio or frequency range of the vertical component of the earthquake.

Deliverable D5.4 outlines the methodology and the topics analysed in T5.4 for deriving PanEuropean observables within the Web-of-Cells (WoC) concept. Observables are derived by analysing traditional instability events typical of large power systems.

This Electra internal report includes the work of Task T6.1 describing the nature, availability and contribution of flexibility resources. This task also models the interactions across control boundaries and identifies sources of control conflicts, giving also an overview of experiences from the ELECTRA partners regarding the realization of controllers in demonstration and field test projects. The work was carried out during the period from May to December 2014. The different type of flexibility resources, their characteristics, affecting market mechanisms and potential for aggregation were researched using a survey among project partners. The parameters used to characterise flexibility include the amount of power modulation, the duration, the rate of change, the response time, the location, the availability, the controllability, etc. Views were also received how these parameters will develop until 2030 and what are the general trends for development of amount and controllability of this resource types. The parameters characterising different energy resources provide the technical requirements for their applicability to flexible operation of the grid and their suitability for frequency and voltage control now and in the future. Regarding the flexibility of electricity generation, gas turbines and other heat motors as reciprocating engines can be started quickest. The speed of power change is clearly the highest for heat motors and their minimum power is low. Also steam and combined heat and power plants can be utilised in the relatively quick increasing of the electricity generation. Slower power changes are possible also with the nuclear power but they cannot be carried out continuously. The regulation characteristics of hydro power are superb in comparison to the other electricity generation methods. Besides the sun power, wind power is increasing most quickly in the world in the comin