1 paper
Qing Li, Yixin Zhu, Yitao Liang +3
Current learning models often struggle with human-like systematic generalization, particularly in learning compositional rules from limited data and extrapolating them to novel com…