V001 / JSI

The Springer publishing house has published a book by associates of the Department of communications systems Arsim Kelmendi, Aleš Švigelj, Tomaž Javornik, and Andrej Hrovat entitled Site Diversity in Satellite Communications: Modelling Using Copula Functions. The book is part of the Springer Briefs series. The book describes multi-site diversity modelling of induced rain attenuation statistics for satellite communication systems using copula functions. It collects all the relevant state-of-the- art findings, provides the missing pieces, and rounds them up in a way that the reader is given a complete picture of the important modelling factors and how to address them. Data post-processing methodology for statistical analysis based on our Earth-satellite propagation experiments are described. Four new models based on Copula functions for predicting diversity at multiple sites are presented with detailed comparative tests and error performance. The results presented in this book are intended to help improve system design and further research into modelling the next generation satellite links at higher frequencies.

Members of the Departments Computer Systems and Artificial Intelligence of Jožef Stefan Institute participate in the Horizon Europe project CONDUCTOR (Fleet and traffic management systems for conducting future cooperative mobility), where JSI develops advanced algorithms and performs technical coordination. As part of the three-year project, advanced traffic and fleet management will be designed, integrated and demonstrated, to allow for efficient and optimal transportation of passengers and goods. Modern algorithms of dynamic balancing and priority-based management of vehicles will be implemented. The project will develop a new generation of models and tools enabled by artificial intelligence and data fusion. Upgraded existing technologies will place autonomous vehicles in the cities of the future, with higher safety and flexible, responsive and centralized mobility control. This will lead to smoother traffic flow, less pollution and a higher quality of life. The resulting open platform will be validated in three use cases in different parts of Europe.

The journal Quantum Machine Intelligence has published an article entitled Boosting the Performance of Quantum Annealers using Machine Learning, authored by researchers from the Department of Knowledge Technologies, the Department of Complex Matter, the Jožef Stefan International Postgraduate School, the Nanocentre and the Faculty of Mathematics and Physics of the University of Ljubljana. The paper is co-authored by researchers Jure Brence, prof. dr. Dragan Mihailović, prof. dr. Viktor V. Kabanov, prof. dr. Ljupčo Todorovski in prof. dr. Sašo Džeroski, under the supervision of Dr. Jaka Vodeb, have managed to minimise the noise in quantum annealers using machine learning methods and thus improve their performance. Quantum annealers are a completely new type of computers that exploit the quantum dynamics of microscopic quantum bits to efficiently solve optimisation problems. They have succeeded in removing the influence of noise to such an extent that they have improved the performance of quantum annealers by three orders of magnitude, opening the way to solving more complex problems.

Researchers from the Department of Knowledge Technologies (E8) and the Computer Systems Department (E7) at JSI recently published a paper entitled OPTION: Optimization algorithm benchmarking ontology in the top-tier journal in Computer Science (IEEE Transactions on Evolutionary Computation, JCR IF = 16.4). The work presented is a collaboration with external partners from Sorbonne University, France, and Leiden University, the Netherlands. The focus of the paper is on the development of an ontology that enables semantic annotation of the core entities involved in the process of benchmarking optimization algorithms (based on evolutionary computation), which allows for a systematic way of sharing and reusing huge amounts of benchmarking data. The development of the OPTION ontology is a step forward in improving the reusability and interoperability of benchmarking data in the domain of optimization. It follows the principles of and contributes to the practice of Open Science.