K2-GNN: Multiple Users’ Comments Integration with Probabilistic K-Hop Knowledge Graph Neural Networks

Huixin Zhan (Texas Tech University); Kun Zhang (XAVIER UNIVERSITY OF LOUISIANA); Chenyi Hu (University of Central Arkansas); Victor S. Sheng (Texas Tech University)*


Integrating multiple comments into a concise statement for any online products or web services requires a non-trivial understanding of the input. Recently, graph neural networks (GNN) has been successfully applied to learn from highly-structured graph representations to mitigate the relationship between entities, such as co-references. However, current inter-sentence relation extraction cannot leverage discrete reasoning chains over multiple comments. To address this issue, in this paper, we propose a probabilistic K-hop knowledge graph (KKG) to extend existing knowledge graphs with inferred relations via discrete intra-sentence and inter-sentence reasoning chains. KKG associates each inferred relation with a confidence value through Bayesian inference. We further answer how a knowledge graph with inferred relations can help the multiple comments integration through integrating KKG with GNN (K2-GNN). Our extensive experimental results show that our K2-GNN outperforms all baseline graph models on multiple comments integration.