ConstGCN: Constrained Transmission-based Graph Convolutional Networks for Document-level Relation Extraction
This addresses a fundamental bottleneck in document-level relation extraction for natural language processing, offering a novel method to improve accuracy.
The paper tackles the graph construction gap in document-level relation extraction by proposing ConstGCN, a graph convolutional network that avoids prior graph construction and uses constrained propagation based on learned transmitting scores, achieving state-of-the-art results on the DocRE dataset.
Document-level relation extraction with graph neural networks faces a fundamental graph construction gap between training and inference - the golden graph structure only available during training, which causes that most methods adopt heuristic or syntactic rules to construct a prior graph as a pseudo proxy. In this paper, we propose $\textbf{ConstGCN}$, a novel graph convolutional network which performs knowledge-based information propagation between entities along with all specific relation spaces without any prior graph construction. Specifically, it updates the entity representation by aggregating information from all other entities along with each relation space, thus modeling the relation-aware spatial information. To control the information flow passing through the indeterminate relation spaces, we propose to constrain the propagation using transmitting scores learned from the Noise Contrastive Estimation between fact triples. Experimental results show that our method outperforms the previous state-of-the-art (SOTA) approaches on the DocRE dataset.