AI Discovers New Physics in the Fourth State of Matter, Challenging Long-Held Assumptions

Physicists at Emory University have demonstrated a groundbreaking use of AI to discover new physics in a dusty plasma, a chaotic system of ionized gas and charged dust particles. By combining a custom neural network with precise 3D particle tracking, the team revealed hidden patterns in non-reciprocal forces, where one particle influences another differently than it is influenced in return. The model achieved over 99% accuracy in describing these forces, which are notoriously difficult to measure and model using traditional methods.

The study, published in PNAS, challenges long-held theoretical assumptions about how these forces behave. The researchers were able to identify and correct inaccuracies in existing models, providing a more accurate understanding of particle dynamics in plasmas. The team, led by Justin Burton and Ilya Nemenman, believes this method is universal and could be applied to other many-body systems, such as industrial materials like paint and ink, or biological systems like groups of living cells. This work was supported by the National Science Foundation and the Simons Foundation.