Extracting Protein-Protein Interactions with Language Modelling

In this paper, we model the corpus-based relation extraction task, namely protein-protein interaction, as a classification problem. In that framework, we first show that standard machine learning systems exploiting representations simply based on shallow linguistic information can rival state-of-the-art systems that rely on deep linguistic analysis. We also show that it is possible to obtain even more effective systems, still using these easy and reliable pieces of information, if the specifics of the extraction task and the data are taken into account. Our original method combining lazy learning and language modelling out-performs the existing systems when evaluated on the LLL2005 protein-protein interaction extraction task data1.