AI in Astrophysics: Domain Shift, Model Robustness & Uncertainty | Aleksandra Ciprijanovic скачать в хорошем качестве

AI in Astrophysics: Domain Shift, Model Robustness & Uncertainty | Aleksandra Ciprijanovic

FirstPrinciples Talks presents 'AI in Astrophysics: Tackling Domain Shift, Model Robustness and Uncertainty' To learn more about FirstPrinciples and the AI Physicist: https://www.firstprinciples.org/ Speaker: Aleksandra Ciprijanovic Artificial intelligence is rapidly transforming astrophysics, from classifying galaxies and studying stellar evolution to mapping cosmic large-scale structure. But as AI moves from controlled benchmarks into real scientific workflows, a critical problem emerges as models trained on simulations or past datasets often fail when applied to new observations. This talk by Aleksandra Ciprijanovic focuses on the deeply connected challenges that determine whether AI can be trusted for discovery in astronomy and cosmology, namely domain shift (when new data differs from training data), model robustness (stability under dataset changes and noise), and uncertainty in predictions (knowing when the model is wrong). Aleksandra highlights real astrophysics examples where these issues appear in practice, including galaxy morphology classification and cosmological parameter inference. The talk also explores strategies to improve model generalization, mitigate bias, and make AI-driven inference more reliable for astrophysics and AI applications across physics, scientific research, and industry. About the speaker Aleksandra is a Wilson Fellow Associate Scientist at the Data Science, Simulation, and Learning Division at Fermi National Accelerator Laboratory, and is also leading the Cosmic Group at the lab. She is interested in the formation and evolution of structures in the Universe - from galaxies and galaxy clusters to large-scale structures. Her work focuses on advancing and building trustworthy and robust AI algorithms that will allow us to fully utilize all available data in the era of large astronomical surveys. Subscribe for more talks on AI-enabled science. #astrophysics #astronomy #machinelearning #aiinscience #cosmology #ai #physics