This repository contains the code and data for the paper "Exploring Pair-Aware Triangular Attention for Biomedical Relation Extraction" accepted by ACM-BCB 2023.
We propose a novel Triangular Attention framework for Biomedical Relation Extraction (called TriA-BioRE) to comprehensively capture pair-aware representations in the biomedical domain. Specifically, we present a triangular attention module, including two triangular multiplications utilizing outgoing and incoming edges, and two triangular self-attention operations centered on the starting and ending nodes, respectively, together to enhance the pair-level modeling omnidirectionally for better BioRE performance.
python>=3.6
pytorch==1.10.2
transformers==4.18.0
numpy==1.19.5
Put the CDR
dataset (including cdr_train.data
, cdr_dev.data
and cdr_test.data
) into folder ./dataset/cdr
.
Put the GDA
dataset (including gda_train.data
, gda_dev.data
and gda_test.data
) into folder ./dataset/gda
.
Put the BioRED
dataset (including biored_train.data
, biored_dev.data
and biored_test.data
) into folder ./dataset/biored
.
python train_triabiore.py