Online Library

A fundamental idea from physics is that macroscopic transitions can occur as a result of an escalation in the correlated activity of a many-body system’s constituent particles. Here we apply this idea in an interdisciplinary setting, whereby the particles are individuals, their correlated activity involves online search activity surrounding the topics of social unrest, and the macroscopic phenomenon being measured are real-world protests. Our empirical study covers countries in Latin America during 2011–2014 using datasets assembled from multiple sources by subject matter experts. We find specifically that the volume and momentum of searches on Google Trends surrounding mass protest language, can detect – and may even pre-empt – the macroscopic on-street activity. Not only can this simple open-source solution prove an invaluable aid for monitoring civil order, our study serves to strengthen the increasing literature in the physics community aimed at understanding the collective dynamics of interacting populations of living objects across the life sciences.

Countering online hate speech is a critical yet challenging task, but one which can be aided by the use of Natural Language Processing (NLP) techniques. Previous research has primarily focused on the development of NLP methods to automatically and effectively detect online hate speech while disregarding further action needed to calm and discourage individuals from using hate speech in the future. In addition, most existing hate speech datasets treat each post as an isolated instance, ignoring the conversational context. In this paper, we propose a novel task of generative hate speech intervention, where the goal is to automatically generate responses to intervene during online conversations that contain hate speech. As a part of this work, we introduce two fully-labeled large-scale hate speech intervention datasets collected from Gab and Reddit. These datasets provide conversation segments, hate speech labels, as well as intervention responses written by Mechanical Turk Workers. In this paper, we also analyze the datasets to understand the common intervention strategies and explore the performance of common automatic response generation methods on these new datasets to provide a benchmark for future research.

Existing work on automated hate speech detection typically focuses on binary classification or on differentiating among a small set of categories. In this paper, we propose a novel method on a fine-grained hate speech classification task, which focuses on differentiating among 40 hate groups of 13 different hate group categories. We first explore the Conditional Variational Autoencoder (CVAE) as a discriminative model and then extend it to a hierarchical architecture to utilize the additional hate category information for more accurate prediction. Experimentally, we show that incorporating the hate category information for training can significantly improve the classification performance and our proposed model outperforms commonly-used discriminative models.

Terrorists and extremists are increasingly utilizing Internet technology to enhance their ability to influence the outside world. Due to the lack of multi-lingual and multimedia terrorist/extremist collections and advanced analytical methodologies, our empirical understanding of their Internet usage is still very limited. To address this research gap, we explore an integrated approach for identifying and collecting terrorist/extremist Web contents. We also propose a Dark Web Attribute System (DWAS) to enable quantitative Dark Web content analysis from three perspectives: technical sophistication, content richness, and Web interactivity. Using the proposed methodology, we identified and examined the Internet usage of major Middle Eastern terrorist/extremist groups. More than 200,000 multimedia Web documents were collected from 86 Middle Eastern multi-lingual terrorist/extremist Web sites. In our comparison of terrorist/extremist Web sites to US government Web sites, we found that terrorists/extremist groups exhibited similar levels of Web knowledge as US government agencies. Moreover, terrorists/extremists had a strong emphasis on multimedia usage and their Web sites employed significantly more sophisticated multimedia technologies than government Web sites. We also found that the terrorists/extremist groups are as effective as the US government agencies in terms of supporting communications and interaction using Web technologies. Advanced Internet-based communication tools such as online forums and chat rooms are used much more frequently in terrorist/extremist Web sites than government Web sites. Based on our case study results, we believe that the DWAS is an effective tool to analyse the technical sophistication of terrorist/extremist groups’ Internet usage and could contribute to an evidence-based understanding of the applications of Web technologies in the global terrorism phenomena.

VOX-Pol Blog post.

VOX-Pol Blog post.