| --- |
| language: |
| - en |
| tags: |
| - politics |
| - roberta |
| license: |
| - cc-by-nc-sa-4.0 |
| --- |
| |
| ## POLITICS |
| POLITICS, a pretrained model on English news articles of politics, is produced via continued training on RoBERTa, based on a **P**retraining **O**bjective **L**everaging **I**nter-article **T**riplet-loss using **I**deological **C**ontent and **S**tory. |
|
|
| **ALERT:** POLITICS is a pre-trained **language model** that specializes in comprehending news articles and understanding ideological content. However, POLITICS cannot be used **out-of-the-box** on downstream tasks such as predicting ideological leanings and discerning stances expressed in texts. To perform predictions on downstream tasks, you are advised to **fine-tune** POLITICS on your own dataset first. |
|
|
| Details of our proposed training objectives (i.e., Ideology-driven Pretraining Objectives) and experimental results of POLITICS can be found in our NAACL-2022 Findings [paper](https://aclanthology.org/2022.findings-naacl.101.pdf) and GitHub [Repo](https://github.com/launchnlp/POLITICS). |
|
|
| Together with POLITICS, we also release our curated large-scale dataset (i.e., BIGNEWS) for pretraining, consisting of more than 3.6M political news articles. This asset can be requested [here](https://forms.gle/yRx5ANHKNuj1kgBDA). |
|
|
| ## Citation |
| Please cite our paper if you use the **POLITICS** model: |
| ``` |
| @inproceedings{liu-etal-2022-POLITICS, |
| title = "POLITICS: Pretraining with Same-story Article Comparison for Ideology Prediction and Stance Detection", |
| author = "Liu, Yujian and |
| Zhang, Xinliang Frederick and |
| Wegsman, David and |
| Beauchamp, Nicholas and |
| Wang, Lu" |
| booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022", |
| year = "2022", |
| ``` |
