Li, F. Compositional Function Networks: A High-Performance Alternative to Deep Neural Networks With Built-in Interpretability. AAAI-SS 2026, 8, 447-454.