The Webis Bias Flipper 2018 (Webis-Bias-Flipper-2018) comprises 2781 events from allsides.com as of June 1st, 2012 till February 10, 2018. For each event, the title, the summary, all news portals belonging to the event, and the links to the news portals with respective bias were recorded. After that, we crawled the news portals with the given links to retrieve their headlines and the content of all articles, because the content is not provided on allsides.com. For each event we collected the corresponding news articles. A total of 6458 news articles are collected.
To download the corpus use the following links:
(12.5 MB, MD5 sum: 3001a4539c8b8e3ea2f9c7bebaca59d0)
If you use the dataset in your research, please send us a copy of your publication. We kindly ask you to refer to the corpus via [bib].
Our research studies the task of “flipping” the bias of news articles: Given an article with a political bias (left or right), generate an article with the same topic but opposite bias. As a first step, we analyze the corpus and discuss intrinsic characteristics of bias. They point to the main challenges of bias flipping, which in turn lead to a specific setting in the generation process. The paper in hand narrows down the general bias flipping task to focus on bias flipping for news article headlines. A manual annotation of headlines from each side reveals that they are self-informative in general and often convey bias. We apply an autoen coder incorporating information from an article’s content to learn how to automatically flip the bias.
For more information on the construction of the dataset see the publication below.
Students: Patrick Saad.