While neural networks offer an attractive way to numerically encode functions. actual formulas remain the language of theoretical particle physics. We use symbolic regression trained on matrix-element information to extract. for instance. https://lecoureures.shop/product-category/ski-de-fond-equipement-batons/