This the the repository for the AAAI 2020 paper "Solving Sequential Text Classification as Board-Game Playing".
GuGo (Game-based jump labeling model) is a jump labeling approach that generalizes the traditional successive labeling by developing a correspondence between sequential text classification and board-game playing.
- code/ contains the source codes.
- data/ contains some datasets used for evaluating.
- Python (≥3.0)
- PyTorch (≥1.0)
- BERT-Base: Please initialize a pretrained BERT model (self.bert in class TextEmbedding) to obtain BERT embeddings.
- Hyperparameters are in _public.py.
If you find this project helps, please kindly consider citing this paper as:
@inproceedings{GuGo,
title = {Solving Sequential Text Classification as Board-Game Playing},
author = {Chen Qian and Fuli Feng and Lijie Wen and Zhenpeng Chen and Li Lin and Yanan Zheng and Tat-Seng Chua},
booktitle = {Proceedings of The 34th AAAI Conference on Artificial Intelligence (AAAI)},
year = {2020}
pages = {8640-8648}
}