Researchers from the University of Ljubljana and the Jožef Stefan Institute, Simon Čopar and Uroš Tkalec, in collaboration with researchers from American universities, reported on the spin ordering of topological defects in a confined nematic liquid crystal that forms under a thin layer of water. Shear forces at the boundary between the liquid crystal and water cause the elastic dipoles to align in the direction of the surface movements and thus store information about external stimuli. This results in ordered domain structures in the nematic, similar to those found in systems with polar order, such as ferromagnets, active matter and metamaterials. This is the first application of a conventional liquid crystal in which the director field is controlled by mechanical motion at an open interface and the defect configurations are read using polarization microscopy. The research, which promises new possibilities for the detection of dynamics in microfluidic environments, was published in Nature Physics. Finally, the article is accompanied by a commentary in the News & Views section. |
In a recent paper in Advanced Materials, Peter Medle Rupnik, Nerea Sebastian and Alenka Mertelj from the Department of Complex Matter and Darja Lisjak from the Material Synthesis Department, in collaboration with researchers from Otto-von-Guericke University, Technical University Braunschweig and Merck Electronics KGaA, have demonstrated an example of a liquid that uniquely exhibits both ferroelectric and ferromagnetic ordering. The breakthrough focuses on a nanostructured liquid crystalline hybrid, composed of ferrimagnetic barium hexaferrite nanoparticles suspended in a ferroelectric nematic host, where director-mediated interactions drive the self-assembly of nanoplatelets in an intricate network. This system shows magnetically driven electric and nonlinear optical responses, alongside electrically driven magnetic responses. This achievement marks significant progress toward the development of multiresponsive multiferroic liquids, with promising potential for advanced applications in energy harvesting, nonlinear optics, and next-generation sensors. |
On July 8-9, 2025, the members of the Natural Language Processing team of the JSI Department of Knowledge Technologies (E8) attended the kick-off meeting of a new European project, ELLIOT, in Thessaloniki (Greece). The project is devoted to training, fine-tuning and applying the next generation of Multimodal Foundation Models, with a commitment to robust, open and fair AI. With a total budget of 25 MIO EUR, this project will run for four years and is executed by a consortium of 30 partners from 12 countries, bringing together leading European AI researchers. With a budget of 700.000 EUR, the JSI team will contribute to foundation model fine-tuning and algorithmic fairness, and will lead the ELLIOT’s community building efforts by supporting the mobility of researchers and organising conferences, workshops, summer schools and other events, in collaboration with the ELIAS and ELLIS scientific consortia. With ELLIOT, Europe positions itself at the frontier of open, trustworthy, and socially beneficial artificial intelligence — capable of driving innovation in critical areas such as media, Earth observation, autonomous systems, and public services. |
The Machine Learning for Physics (ML4Physics) Summer School successfully concluded last week in Ljubljana, Slovenia, bringing together 55 participants — 47 from abroad — for a week-long deep dive into the intersection of machine learning and physics. The school was organized by the COMETA COST Action, SMARTHEP, the SMASH MSCA COFUND, the Jožef Stefan Institute, and the Faculty of Mathematics and Physics at the University of Ljubljana. Its goal was to equip participants from diverse backgrounds with solid knowledge of cutting-edge machine learning topics relevant to physics. Each day combined morning lectures, afternoon hands-on exercises, and evening keynote talks or social and outreach activities. The school featured 11 lecturers and tutors, as well as two keynote talks by Uroš Seljak (UC Berkeley) and Nenad Tomasev (DeepMind), who shared insights from both academia and industry. The ML4Physics Summer School proved to be a vibrant platform for learning, collaboration, and exchange, demonstrating the power of interdisciplinary training in shaping the future of AI in science. |
