Learning Informative Representations of Biomedical Relations with Latent Variable Models

Abstract

Extracting biomedical relations from large corpora of scientific documents is a challenging natural language processing task.  Existing approaches usually focus on identifying a relation either in a single sentence (mention-level)or across an entire corpus (pair-level). In both cases, recent methods have achieved strong results by learning a point estimate to represent the relation;  this is then used as the input to a relation classifier.  However, the relation expressed in the text between a  pair of biomedical entities is often more complex than can be captured by a point estimate.  To address this issue, we propose a latent variable model with an arbitrarily flexible distribution to represent the relation between an entity pair.  Additionally, our model provides a unified architecture for both mention-level and pair-level relation extraction.  We demonstrate that our model achieves results competitive with strong base-lines for both tasks while having fewer parameters and being significantly faster to train. We make our code publicly available.

Github

The code can be accessed on GitHub here.