V001 / JSI

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.

The researchers from Department F6 have achieved an ultrafast differentiation by genomic GC content of the bacterial DNA at a molecular level using surface-enhanced Raman spectroscopy (SERS) and fine-tuned plasmonic gold nanocrystals. Their atmospheric-pressure plasma-designed truncated plasmonic particulates revealed truly exceptional optical response within nanometric gaps between gold dimers/chains, from where the molecular fingerprints of bacterial DNA fragments gained the strongest Raman signal enhancement. It allowed not only to collect vibrational data from different DNA samples but also, for the first time, to distinguish reliably between bacterial species due to intrinsically different compositions of nucleobases adenine, guanine, cytosine, and thymine within the biological macromolecule. The realized approach was successfully validated by cutting-edge third-generation sequencing nanopore technology and published in NanoLetters.