New technique overcomes spurious correlations problem in AI
Peer-Reviewed Publication
Updates every hour. Last Updated: 29-Apr-2025 00:08 ET (29-Apr-2025 04:08 GMT/UTC)
AI models often rely on “spurious correlations,” making decisions based on unimportant and potentially misleading information. Researchers have now discovered these learned spurious correlations can be traced to a very small subset of the training data and have demonstrated a technique that overcomes the problem.
EPFL researchers have achieved a remarkable result: capturing and studying phase changes in quantum hardware, which hold hold promise for next-generation technologies like quantum computing and ultra-sensitive sensors.
At ultracold temperatures, interatomic collisions are relatively simple, and their outcome can be controlled using a magnetic field. However, research by scientists led by Prof. Michal Tomza from the Faculty of Physics of the University of Warsaw and prof. Roee Ozeri from the Weizmann Institute of Science shows that this is also possible at higher temperatures. The scientists published their observations in the scientific journal Science Advances.