AI Summary: An international team led by Aalto University Professor Päivi Törmä has demonstrated that machine learning can significantly expedite the search for new superconductors, materials that conduct electricity without resistance. The SuperC consortium utilized AI to screen vast combinations of elemental materials, identifying promising candidates that were further analyzed through quantum calculations. This approach led to the successful synthesis and verification of two new superconductors, YRu3B2 and LuRu3B2, which exhibit superconducting properties due to their unique electronic structures. The findings, published in *Physical Review Research*, aim to facilitate the discovery of room-temperature superconductors, potentially transforming energy consumption in various technologies.
Aurora 1.5: Extending open foundation models for weather and Earth-system applications
AI Summary: Aurora 1.5 is an updated version of Microsoft's Aurora Earth System foundation model, which now includes 22 additional weather variables, enhances temporal resolution to hourly forecasts, and introduces probabilistic ensemble forecasting. Released as open source on GitHub and with model checkpoints on Hugging Face, it allows researchers and developers to evaluate and build upon the model. This extension aims to improve operational guidance for sectors such as energy, agriculture, and transport by providing a more comprehensive view of atmospheric conditions and supporting decision-making in the face of climate risks. The update reflects a commitment to making advanced weather forecasting tools more accessible and practical for various organizations.