This repository contains code for Multi-Task Learning for Relation Extraction in Proceedings of 2019 IEEE 31st International Conference on Tools with Artificial Intelligence (ICTAI).
This code has been run with TensorFlow 1.14, TensorPack 0.9.7 and Numpy 1.16.3; other versions may work, but have not been tested.
See workflow.pdf for detail. We use Riedel 2010 dataset for evaluation. For part of dependency and entity type labels, we thank RESIDE for providing processed data on their github page. For other missing labels, we use StanfordNLP to obtain dependency labels and FIGER to obtain entity type labels.
Run python edr.py pretrain for pretraining, python edr.py train for training and python edr.py eval for evaluation.