Identifying Protein-Protein Interaction using Tree LSTM and Structured Attention

TL;DR

This paper introduces a novel tree LSTM with structured attention for protein-protein interaction extraction, achieving state-of-the-art results without handcrafted features, highlighting the effectiveness of tree-based neural networks.

Contribution

The paper presents a new tree LSTM architecture with structured attention that outperforms previous models in PPI extraction without manual feature engineering.

Findings

Achieves state-of-the-art precision, recall, and F1-score on benchmark datasets.

Tree recurrent networks outperform traditional recurrent networks in PPI tasks.

Significant improvement over previous models without explicit feature extraction.

Abstract

Identifying interactions between proteins is important to understand underlying biological processes. Extracting a protein-protein interaction (PPI) from the raw text is often very difficult. Previous supervised learning methods have used handcrafted features on human-annotated data sets. In this paper, we propose a novel tree recurrent neural network with structured attention architecture for doing PPI. Our architecture achieves state of the art results (precision, recall, and F1-score) on the AIMed and BioInfer benchmark data sets. Moreover, our models achieve a significant improvement over previous best models without any explicit feature extraction. Our experimental results show that traditional recurrent networks have inferior performance compared to tree recurrent networks for the supervised PPI problem.