Search for: All records

The ideological asymmetries have been recently observed in contested online spaces, where conservative voices seem to be relatively more pronounced even though liberals are known to have the population advantage on digital platforms. Most prior research, however, focused on either one single platform or one single political topic. Whether an ideological group garners more attention across platforms and/or topics, and how the attention dynamics evolve over time, have not been explored. In this work, we present a quantitative study that links collective attention across two social platforms -- YouTube and Twitter, centered on online activities surrounding popular videos of three controversial political topics including Abortion, Gun control, and Black Lives Matter over 16 months. We propose several sets of video-centric metrics to characterize how online attention is accumulated for different ideological groups. We find that neither side is on a winning streak: left-leaning videos are overall more viewed, more engaging, but less tweeted than right-leaning videos. The attention time series unfold quicker for left-leaning videos, but span a longer time for right-leaning videos. Network analysis on the early adopters and tweet cascades show that the information diffusion for left-leaning videos tends to involve centralized actors; while that for right-leaning videos starts earlier in the attention lifecycle. In sum, our findings go beyond the static picture of ideological asymmetries in digital spaces and provide a set of methods to quantify attention dynamics across different social platforms. 

The increasing and flexible use of autonomous systems in many domains -- from intelligent transportation systems, information systems, to business transaction management -- has led to challenges in understanding the normal and abnormal behaviors of those systems. As the systems may be composed of internal states and relationships among sub-systems, it requires not only warning users to anomalous situations but also provides transparency about how the anomalies deviate from normalcy for more appropriate intervention. We propose a unified anomaly discovery framework DeepSphere that simultaneously meet the above two requirements -- identifying the anomalous cases and further exploring the cases' anomalous structure localized in spatial and temporal context. DeepSphere leverages deep autoencoders and hypersphere learning methods, having the capability of isolating anomaly pollution and reconstructing normal behaviors. DeepSphere does not rely on human annotated samples and can generalize to unseen data. Extensive experiments on both synthetic and real datasets demonstrate the consistent and robust performance of the proposed method.