Abstract
Inter-sentence relation extraction deals with a number of complex semanticrelationships in documents, which require local, non-local, syntactic andsemantic dependencies. Existing methods do not fully exploit such dependencies.We present a novel inter-sentence relation extraction model that builds alabelled edge graph convolutional neural network model on a document-levelgraph. The graph is constructed using various inter- and intra-sentencedependencies to capture local and non-local dependency information. In order topredict the relation of an entity pair, we utilise multi-instance learning withbi-affine pairwise scoring. Experimental results show that our model achievescomparable performance to the state-of-the-art neural models on twobiochemistry datasets. Our analysis shows that all the types in the graph areeffective for inter-sentence relation extraction.