L. Tian, X. Zhang, and J.H. Lau (2021). Rumour Detection via Zero-shot Cross-lingual Transfer Learning. In Proceedings of The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2021), pages 603—618.
We use Twitter15/Twitter16, PHEME and WEIBO datasets for rumour detection. The five-fold split is from here
In this repository, we do not provide you with the raw input data. Please download the datasets from the following links.
| Dataset | Link |
|---|---|
| T15/T16 | https://www.dropbox.com/s/7ewzdrbelpmrnxu/rumdetect2017.zip?dl=0 ) |
| PHEME | https://figshare.com/articles/dataset/PHEME_dataset_of_rumours_and_non-rumours/4010619 |
| https://alt.qcri.org/~wgao/data/rumdect.zip | |
| Extra Chinese Pretraining Data | https://archive.ics.uci.edu/ml/datasets/microblogPCU |
| Extra English Pretraining Data | https://github.com/KaiDMML/FakeNewsNet |
- Python 3.6
- Run
pip install -r requirements.txt
Clone the transformer to local directory
git clone https://github.com/huggingface/transformers.git
For further pre-training the language models:
python transformers/examples/language-modeling/run_language_modeling.py ,
--output_dir='ML_MBERT_DAPT',
--model_type=bert ,
--model_name_or_path=bert-base-multilingual-cased-freeze-we,
--do_train,
--overwrite_output_dir,
--train_data_file='train.txt',
--do_eval,
--block_size=256,
--eval_data_file='vali.txt',
--mlm"
#If you find this code useful, please let us know and cite our paper.