Symbolic regression as AI-based tool for the analysis of smartphone experiments
M. Ziese · Physics Education · 2025
Abstract Symbolic regression (SR) is discussed as an artificial intelligence-based tool to derive the mathematical expressions of physical laws directly from data that can be easily measured in a physics lab in universities or in schools. As an example the acceleration of a rotating iPad is analysed in terms of the angular velocity and angular acceleration. SR algorithms consistently find the correct functional expressions and also provide the coordinates of the accelerometer with respect to the centre of the iPad. This tool opens up a new approach to learning in physics, where data is the primary source that lead to physical laws, as opposed to the standard procedure of memorizing laws and validating these through experiments.