Ethical issues, along with transparency, disinformation and bias are in the focus of our information society. In this demo, we will present the bias goggles system, that based on the web graph computes the bias characteristics of web domains to user-defined concepts. Our approach uses adaptations of propagation models and a variation of the pagerank algorithm named Biased-PR, that models various behaviours of biased surfers. Currently, the system runs over a subset of the greek web graph. We have developed cross-browser plugins that let users explore the bias characteristics of domains, define their own biased concepts and monitor the consumption of biased information through some analytics.

Details #

• Publication: In International Workshop on Algorithmic Bias in Search and Recommendation
• DOI: https://doi.org/10.1007/978-3-030-52485-2_18
• Code: https://github.com/probonas/bias-goggles-web-extension
• Slides: http://publications.ics.forth.gr/_publications/Papadakos.pdf