SoREX: Towards Self-Explainable Social Recommendation with Relevant Ego-Path Extraction
Published in ACM Transactions on Information Systems (TOIS), CCF-A, 2025
SoREX introduces a self-explainable GNN-based framework for social recommendation. It models social relations and user-item interactions with a dual-tower design, extracts factor-specific ego-paths, and re-aggregates explanations to provide fine-grained and traceable recommendation evidence.
| ACM DL | Code |
Recommended citation: Hanze Guo, Yijun Ma, and Xiao Zhou. "SoREX: Towards Self-Explainable Social Recommendation with Relevant Ego-Path Extraction." ACM Transactions on Information Systems, 2025.
Download Paper