V001 / JSI

AutoLearn-SI is a €2.5M #HorizonEU-funded project at Jožef Stefan Institute, establishing a top-tier ERA Chair research group to advance automated machine learning and optimization (AutoML and AutoOPT). It integrates cutting-edge AutoML/AutoOPT into research and education, supporting national reforms and boosting EU funding potential. The project is built on three core vertical pillars - Experimental Databases, Representation Learning, and Automated Algorithm Selection/Configuration - alongside two horizontal pillars: Single-objective Optimization (SOO) and Multi-label Classification (MLC). Led by a USA-based ERA Chair Holder, AutoLearn-SI fosters international collaboration with partners from France, Germany, the Netherlands, and Belgium while enabling knowledge transfer to North Macedonia and other widening countries. By strengthening research capacity and stakeholder engagement within the European Research Area, the project drives innovation, enhances Slovenia’s leadership in automated machine learning and optimization, and contributes to regional excellence in AI-driven decision-making.

In partnership with the Faculty of Engineering, University of Ljubljana, Department of Nanostructured Materials wes awarded the third Rector's Award for the patent as a final result of the EIT Manufacturing project aProMag. The main objective of the project was to prototype, validate and bring to market a technology for rapid prototyping of rotors for brushless DC motors and actuators using 3D printing in a magnetic field that enables anisotropic alignment of hard magnetic material. The advanced technology reduces waste through 3D printing technology, with very low waste and material used reusable up to 5 times (~97% material yield). Advanced technology reduces waste through 3D printing technology. The source of feedstock powder is raw material recovered from end-of-life NdFeB magnets, enabling a circular economy. The technology also significantly reduces the time to produce end products to test design concepts by up to 5 times and significantly reduces the cost of conventional injection moulding tooling. The end user to test the new methods is Kolektor, a world-renowned manufacturer in the field. The technology is patented.

Dr Jaka Vodeb from Department of Complex Matter at the Jožef Stefan Institute has published a study in Nature Physics in which he has used quantum simulation to gain valuable insight into a well-known physical phenomenon, the false vacuum decay. Together with colleagues from the University of Leeds, the Forschungszentrum Jülich, where Dr Vodeb did his postdoctoral education, and the Austrian Institute of Science and Technology (ISTA), they sought to understand the key puzzle of the false vacuum decay and the mechanism behind it. The experiment involved placing 5564 qubits - the basic building blocks of quantum computing - in specific configurations to represent a false vacuum. By carefully controlling the system, they were able to trigger a transition from a false vacuum to a real vacuum, mirroring the formation of bubbles as described by the theory of false vacuum decay. Although this process could trigger a significant change in the structure of the Universe, predicting its timing is difficult; it is likely to occur over a period that could last millions or even billions of years.

An international team of scientists published an article First-order quantum breakdown of superconductivity in an amorphous superconductorin the journal Nature Physics. Mikhail Feigel’man from Jožef Stefan Institute (Dept. of Complex Matter F7) and the CENN Nanocenter, in collaboration with researchers from the Neel Institute (CNRS Grenoble) and Karlsruhe Institute of Technology, provided a groundbreaking framework for understanding of unexpected experimental demonstration of sharp disappearance of superconductivity of thin films at ultra-low temperatures, upon increase of their normal-state resistance, that is in sharp contrast with standard paradigm of continuous (second-order) superconducting-insulator transitions existing for more than three decade. The first-order nature of this transition is understood in terms of energy competition between two low-temperature states of matter: the superconducting condensate of electron pairs on one hand, and the insulator made out of bound electron pairs with Coulomb repulsion between them, on another hand.