AAAI20 "CopyMTL: Copy Mechanism for Joint Extraction of Entities and Relations with Multi-Task Learning"
Paper accepted by AAAI-2020
This is a followup paper of "Extracting Relational Facts by an End-to-End Neural Model with Copy Mechanism" ACL2018 CopyRE
python3
pytorch 0.4.0 -- 1.3.1
This repo initially contain webnlg, you can run the code directly. NYT dataset need to be downloaded and to be placed in proper path. see const.py.
The pre-processed data is avaliable in:
WebNLG dataset: https://drive.google.com/open?id=1zISxYa-8ROe2Zv8iRc82jY9QsQrfY1Vj
NYT dataset: https://drive.google.com/open?id=10f24s9gM7NdyO3z5OqQxJgYud4NnCJg3
python main.py --gpu 0 --mode train --cell lstm --decoder_type one
python main.py --gpu 0 --mode test --cell lstm --decoder_type one