Distant Supervision for Relation Extraction with Multi-sense Word Embedding

Abstract

Distant supervision can automatically generate labeled data between a large-scale corpus and a knowledge base without utilizing human efforts. Therefore, many studies have used the distant supervision approach in relation extraction tasks. However, existing studies have a disad-vantage in that they do not reflect the homo-graph in the word embedding used as an input of the relation extraction model. Thus, it can be seen that the relation extraction model learns without grasping the meaning of the word ac-curately. In this paper, we propose a relation ex-traction model with multi-sense word embed-ding. We learn multi-sense word embedding using a word sense disambiguation module. In addition, we use convolutional neural network and piecewise max pooling convolutional neural network relation extraction models that effi-ciently grasp key features in sentences. To eval-uate the performance of the proposed model, two additional methods of word embedding were learned and compared. Accordingly, our method showed the highest performance among them.