Additional context helps! Leveraging cited paper information to improve citation classification

With the rapid growth in research publications, automated solutions to tackle scholarly information overload is growing more relevant. Correctly identifying the intent of the citations is one such task that finds applications ranging from predicting scholarly impact, finding idea propagation, to text summarization to establishing more informative citation indexers. In this in-progress work, we leverage the cited paper's information and demonstrate that this helps in the effective classification of citation intents. We propose a neural multi-task learning framework that harnesses the structural information of the research papers and the relation between the citation context and the cited paper for citation classification. Our initial experiments on three benchmark citation classification datasets show that with incorporating cited paper information (title), our neural model achieves a new state of the art on the ACL-ARC dataset with an absolute increase of 5.3% in the F1 score over the previous best model. Our approach also outperforms the submissions made in the 3C Shared task: Citation Context Classification with an increase of 8% and 3.6% over the previous best Public F1-macro and Private F1-macro scores respectively.