Darwin's Theory of Censorship: Analysing the Evolution of Censored Topics with Dynamic Topic Models

What is it about?

The paper presents a technique to detect how censorship policies evolve in reaction to real events without prior knowledge of what these events are. Using longitudinal data collected from ICLab, together with historical records of web pages collected by the Internet Archive, we validate our technique using censored websites in India. We find that gaps in data collection severely impact the analysis, and present recommendations to further facilitate such analysis.

Featured Image

Why is it important?

While most work has focused on finding topics of censorship in a single snapshot of time, our work focuses on capturing the evolution of censorship instead. This gives us more nuanced insight into how censorship policies change with respect to significant events. Furthermore, it also allows us to infer how the discussions on censored topics evolve. For example, while censored content would initially be about a specific event, the same topic can then evolve into more general discussions around freedoms and rights.