The paper introduces FedCQA, a method to answer complex queries on multi-source knowledge graphs (KGs) while preserving privacy. Current methods focus on single KGs and can’t be applied to multiple graphs. Moreover, sharing KGs with sensitive info can lead to privacy risks. Hence, federated learning is used to collaboratively learn representations with privacy preserved. FedCQA enhances the relations in KGs, improving representation quality. Extensive experiments show that FedCQA retrieves answers to cross knowledge graph queries while keeping raw data secret.


Publication date: 23 Feb 2024
Project Page: Not provided