Construction of Bilingual Corpus Knowledge Graph and Semantic Association Mining Fused with GNN

Aiming at the complexity of knowledge representation and semantic association in bilingual corpora, a method for constructing bilingual KG and mining semantic association fused with GNN is proposed. First, semantic resources of different languages are integrated through multilingual data preprocessing and fusion technologies; second, entity matching of bilingual KG is realized by using a GNN-based cross-lingual entity alignment algorithm; finally, in-depth cross-lingual semantic relationships are revealed through the semantic association mining module. Experimental results show that the method proposed in this paper achieves an accuracy of 89.6% in the cross-lingual entity alignment task, an increase of 12.4% compared with traditional methods, and can effectively mine potential semantic associations in bilingual corpora. This research provides an innovative solution for the construction of cross-lingual KG and has important value for applications such as machine translation and cross-lingual information retrieval.