Tag
EditSR proposes a two-layer framework combining a neural symbolic regression model with an edit-based Rectifier to efficiently rectify structural errors in generated expressions, reducing error accumulation and improving recovery of complex symbolic structures with limited extra cost.