Three-Branch Decision-Guided Multi-Granularity Knowledge Reasoning with Graph Neural Networks

With the widespread application of Graph Neural Networks (GNNs) in knowledge graph reasoning tasks, existing methods still face challenges in terms of reasoning accuracy and efficiency, especially when dealing with complex and sparse knowledge graphs, where model performance often suffers. To address this issue, this paper proposes a Three-Branch Decision-Guided Multi-Granularity Knowledge Reasoning framework for Graph Neural Networks, which combines a three-branch decision guidance mechanism with a multi-granularity reasoning strategy to enhance the reasoning accuracy and computational efficiency of the model. Through experiments on multiple benchmark datasets such as FB15k, WN18, and NELL, we validate the superiority of the proposed method in terms of accuracy, recall, and F1-score, while also demonstrating high efficiency in reasoning speed. Experimental results show that the three-branch decision guidance mechanism and multi-granularity reasoning strategy effectively improve the model's reasoning performance, verifying the effectiveness of the proposed framework. This research provides new ideas and approaches for the application of Graph Neural Networks in the field of knowledge reasoning, playing a significant role in advancing both theoretical research and practical applications, such as intelligent question answering and recommendation systems.