Perspectrum

This repository contains codes and scripts relevant to the dataset introduced in this paper.

Code structure:

• website/: contains the webapp used for crowdsourcing experiments, as well as the search engine for claim/perspectives.
• data/: contains the dataset at each stages of our annotations.
• experiments: Jupyter notebooks containing different small experiments for dataset construction.

Model

You can download the BERT-based baseline models for all four tasks from the google drive.

Demo

We host a demo of our BERT-based baseline systems here The server code for the demo is available at this repo

Citation

Work is published under the Creative Commons license. Please cite the following work if you want to refer to this work:

@inproceedings{chen2018perspectives,
  title={Seeing Things from a Different Angle: Discovering Diverse Perspectives about Claims},
  author={Chen, Sihao and Khashabi, Daniel and Yin, Wenpeng and Callison-Burch, Chris and Roth, Dan},
  book={Conference of the North American Chapter of the Association for Computational Linguistics (NAACL)},
  year={2019}
}