F. Li, “Compositional Function Networks: A High-Performance Alternative to Deep Neural Networks with Built-in Interpretability”, AAAI-SS, vol. 8, no. 1, pp. 447–454, May 2026.