Search

In this paper we present a near-complete dataset of over 3M videos from 61K channels over 2.5 years (June 2019 to December 2021) from the social video hosting platform BitChute, a commonly used alternative to YouTube. Additionally, we include a variety of video-level metadata, including comments, channel descriptions, and views for each video. The MeLa-BitChute dataset can be found at: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/KRD1VS.

This paper investigates visual media shared by US national politicians on Twitter, how a politician's variety of image types shared reflects their political position, and identifies a hazard in using standard methods for image characterization in this context. While past work has yielded valuable results on politicians' use of imagery in social media, that work has focused primarily on photographic media, which may not be sufficient given the variety of visual media shared in such spaces (e.g., infographics, illustrations, or memes). Leveraging three popular deep learning models to characterize politicians' visuals, this work uses clustering to identify eight types of visual media shared on Twitter, several of which are not photographic in nature. Results also show individual politicians share a variety of these types, and the distributions of their imagery across these clusters is correlated with their overall ideological position -- e.g., liberal politicians appear to share a larger proportion of infographic-style images, and conservative politicians appear to share more patriotic imagery. At the same time, manual assessment reveals that these image characterization models group images with vastly different semantic meaning into the same clusters, as confirmed in a post-hoc analysis of hateful memetic imagery. These results suggest that, while image-characterization techniques do identify general types of imagery that correlate with political ideology, these methods miss critical semantic -- and therefore politically relevant -- differences among images. Consequently, care should be taken when dealing with the varieties of imagery shared in online spaces, especially in political contexts.

In this paper, we characterize the content and discourse on BitChute, a social video-hosting platform. Launched in 2017 as an alternative to YouTube, BitChute joins an ecosystem of alternative, low content moderation platforms, including Gab, Voat, Minds, and 4chan. Uniquely, BitChute is the first of these alternative platforms to focus on video content and is growing in popularity. Our analysis reveals several key characteristics of the platform. We find that only a handful of channels receive any engagement, and almost all of those channels contain conspiracies or hate speech. This high rate of hate speech on the platform as a whole, much of which is anti-Semitic, is particularly concerning. Our results suggest that BitChute has a higher rate of hate speech than Gab but less than 4chan. Lastly, we find that while some BitChute content producers have been banned from other platforms, many maintain profiles on mainstream social media platforms, particularly YouTube. This paper contributes a first look at the content and discourse on BitChute and provides a building block for future research on low content moderation platforms., This long paper is supplemental to a short version of the paper published in ACM Conference on Hypertext and Social Media 2020

The ubiquity of mobile internet-enabled devices combined with wide-spread social media use during emergencies is posing new challenges for response personnel. In particular, service operators are now expected to monitor these online channels to extract actionable insights and answer questions from the public. A lack of adequate tools makes this monitoring impractical at the scale of many emergencies. The TREC Incident Streams (TREC-IS) track drives research into solving this technology gap by bringing together academia and industry to develop techniques for extracting actionable insights from social media streams during emergencies. This paper covers the second year of TREC-IS, hosted in 2019 with two editions, 2019-A and 2019-B, contributing 12 new events and approximately 20,000 new tweets across 25 information categories, with 15 research groups participating across the world. This paper provides an overview of these new editions, actionable insights from data labelling, and the automated techniques employed by participant systems that appear most effective.

The Text Retrieval Conference (TREC) Incident Streams track is a new initiative that aims to mature social\ud media-based emergency response technology. This initiative advances the state of the art in this area through an\ud evaluation challenge, which attracts researchers and developers from across the globe. The 2018 edition of the track\ud provides a standardized evaluation methodology, an ontology of emergency-relevant social media information types,\ud proposes a scale for information criticality, and releases a dataset containing fifteen test events and approximately\ud 20,000 labeled tweets. Analysis of this dataset reveals a significant amount of actionable information on social\ud media during emergencies (> 10%). While this data is valuable for emergency response efforts, analysis of the\ud 39 state-of-the-art systems demonstrate a performance gap in identifying this data. We therefore find the current\ud state-of-the-art is insufficient for emergency responders’ requirements, particularly for rare actionable information\ud for which there is little prior training data available.