AI Unearths Hidden Physics in Dusty Plasma Through Machine Learning

Physicists have turned to artificial intelligence to uncover unexpected new behaviors in dusty plasma, a complex many-body system made up of charged microparticles suspended in a gas. Using a machine-learning approach, researchers identified previously unknown non-reciprocal forces—interactions where the force exerted by one particle on another does not equal the reverse—challenging conventional assumptions in classical physics. Dusty plasmas, found in space environments like planetary rings and interstellar clouds, as well as in laboratory settings, are ideal for studying collective behavior in systems far from equilibrium. In these systems, particles interact through electrostatic forces, but their motion is also influenced by external fields and collisions with surrounding gas molecules. While some non-reciprocal interactions have been observed before, the full complexity of these forces in large, dynamic systems has remained poorly understood. The team applied a deep learning algorithm trained on high-resolution particle trajectory data collected from experiments. Instead of relying on pre-defined physical models, the AI was allowed to detect patterns and correlations directly from the data. To their surprise, the model revealed that certain interactions between particles were not symmetric—meaning the force from particle A on particle B was significantly different from the force from B on A—under specific conditions. These non-reciprocal forces were not predicted by existing theories and appear to emerge from the collective dynamics of the system, particularly when particles are driven out of equilibrium by external electric fields. The findings suggest that traditional physics frameworks may need to be extended to account for such behaviors in complex, real-world systems. The discovery highlights the power of machine learning as a tool for scientific exploration, especially in domains where analytical models are too limited to capture emergent phenomena. Rather than replacing established theory, the AI acted as a discovery engine, guiding physicists toward new physical insights that would have been difficult to detect through conventional analysis. Researchers say the results could have implications beyond dusty plasmas, potentially influencing fields such as soft matter physics, active matter, and even the design of new materials with adaptive or self-organizing properties. As AI continues to integrate with experimental science, it may become a key partner in uncovering the hidden rules that govern complex systems.