This repository includes the code of the paper Optimizing attention for sequence modeling via reinforcement learning published at IEEE TNNLS.
python>=3.6
tensorflow
tflearn
numpyText classifications:
- Movie Review (MR)
- AGnews
- Subjectivity (SUBJ)
- Stanford Sentiment Treebank (SST)
prepare the glove word embedding at emb file:
glove.6B.100d.txt
python core/main.pyIf you use this work, please kindly cite:
@article{FeiZRJ22,
author = {Hao Fei and
Yue Zhang and
Yafeng Ren and
Donghong Ji},
title = {Optimizing Attention for Sequence Modeling via Reinforcement Learning},
journal = {{IEEE} Trans. Neural Networks Learn. Syst.},
volume = {33},
number = {8},
pages = {3612--3621},
year = {2022},
url = {https://doi.org/10.1109/TNNLS.2021.3053633}
}
The code is released under Apache License 2.0 for Noncommercial use only.