AI Uncovers New Physical Laws in Dusty Plasma Phenomena

Researchers at Emory University have achieved a significant breakthrough by employing machine learning to uncover novel physical principles governing dusty plasma—a distinctive form of matter present in space and natural phenomena like wildfires. This discovery not only enhances our understanding of particle interactions within this enigmatic plasma but also demonstrates AI's potential to discover new laws of physics beyond data analysis and predictions. Published in the Proceedings of the National Academy of Sciences, the study was led by physicists Justin Burton and Ilya Nemenman. Unlike traditional AI models treated as opaque "black boxes," the team designed a transparent system that allowed them to understand how and why specific results emerged. Dusty plasma comprises ionized gases mixed with charged particles such as dust or soot. It exists throughout the universe—from Saturn's rings to lunar surfaces—and on Earth, it manifests during wildfires, where charged soot particles can disrupt radio signals. The research involved suspending plastic particles in a plasma-filled vacuum chamber, illuminated by a laser, with high-speed cameras capturing their three-dimensional movemen...