Evolutionary Identification of Scientific Equations Using Language-Model-Guided Symbolic Regression
Results show that incorporating language-model guidance into evolutionary SR substantially improves search efficiency while preserving interpretability and the evolutionary character of the algorithm.