Bias Bubbles: Using Semi-Supervised Learning to Measure How Many Biased News Articles Are Around Us

Qin Ruan, Brian Mac Namee, Ruihai Dong
2021 Irish Conference on Artificial Intelligence and Cognitive Science  
The proliferation of web 2.0 technology allows us to easily create and share online content, but also leads to the rapid spread of misinformation and biased media, which has considerable negative effects on society. Deep learning-based classifiers are one common way of identifying media bias, but they suffer from a lack of large-scale labelled datasets. In this paper, we first explore the use of pseudo-labelling technology to mitigate this problem. Second, we exploit a masking method to
more » ... biased sentences in news articles by iteratively masking each sentence from an article and observing the change in output of a bias detection model. These identified sentences not only contribute to evaluating the proposed model, but also enable end-users to understand where media bias arises in an article. Finally, we apply our well-trained bias detection model to a well-known news article dataset to show how widespread media bias is-the results show that it is rampant and has become a serious social problem that we cannot ignore.
dblp:conf/aics/RuanND21 fatcat:d5za3marpbg5dpcoycb3nzfray