Towards human-like perception: Learning structural causal model in heterogeneous graph

Tianqianjin Lin, Kaisong Song, Zhuoren Jiang, Yangyang Kang, Weikang Yuan, Xurui Li, Changlong Sun, Cui Huang, Xiaozhong Liu

Abstract: Highlights•This work proposes a novel heterogeneous graph algorithm, namely HG-SCM.•HG-SCM aligns with human reasoning logic by introducing structural causal model.•HG-SCM excels in both performance and generalizability compared to existing methods.•HG-SCM has task-level interpretability that is easily understandable for humans.