Trends in Detection and Characterization of Propaganda Bots

Williamson, W.; Scrofani, James

doi:10.24251/hicss.2019.854

Cited by 10 publications

(7 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BERT (Zellers et al, 2019), RoBERTa and ELMo (Cruz et al, 2019)) have been proposed to handle this fragment-level task. To detect propaganda on social media, many studies (Williamson III and Scrofani, 2019;Caldarelli et al, 2020) utilize datasets that contain propaganda content spread on social media by Russian-based IRA (Farkas and Bastos, 2018;Miller, 2019) or extremists (Johnston and Weiss, 2017;Nizzoli et al, 2019). Subsequent works investigated the influence of propaganda on public opinion (Caldarelli et al, 2020) and the techniques applied to disseminate targeted political agenda (Gorrell et al, 2019).…”

Section: Related Workmentioning

confidence: 99%

TWEETSPIN: Fine-grained Propaganda Detection in Social Media Using Multi-View Representations

Vijayaraghavan¹,

Vosoughi²

2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

Recently, several studies on propaganda detection have involved document and fragmentlevel analyses of news articles. However, there are significant data and modeling challenges dealing with fine-grained detection of propaganda on social media. In this work, we present TWEETSPIN, a dataset containing tweets that are weakly annotated with different fine-grained propaganda techniques, and propose a neural approach to detect and categorize propaganda tweets across those fine-grained categories. These categories include specific rhetorical and psychological techniques, ranging from leveraging emotions to using logical fallacies. Our model relies on multi-view representations of the input tweet data to (a) extract different aspects of the input text including the context, entities, their relationships, and external knowledge; (b) model their mutual interplay; and (c) effectively speed up the learning process by requiring fewer training examples. Our method allows for representation enrichment leading to better detection and categorization of propaganda on social media. We verify the effectiveness of our proposed method on TWEETSPIN and further probe how the implicit relations between the views impact the performance. Our experiments show that our model is able to outperform several benchmark methods and transfer the knowledge to relatively low-resource news domains.

show abstract

Section: Related Workmentioning

confidence: 99%

TWEETSPIN: Fine-grained Propaganda Detection in Social Media Using Multi-View Representations

Vijayaraghavan¹,

Vosoughi²

2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

show abstract

“…Substantial academic effort has been put into the detection of fake accounts (e.g. Davis et al, 2016; Varol et al, 2017; Williamson and Scrofani, 2019) and false information diffusion (e.g. Bastos and Mercea, 2019; Grinberg et al, 2019; Shin et al, 2017; Vosoughi et al, 2018) by analyzing large bodies of digital data.…”

Section: Literature Reviewmentioning

confidence: 99%

Fake thumbs in play: A large-scale exploration of false amplification and false diminution in online news comment spaces

Kwon

Lee

Han

et al. 2022

New Media & Society

View full text Add to dashboard Cite

This study explores how disinformation can dampen general users’ expressions of opinion online. In the context of a proven disinformation case in South Korea, this study analyzes externally validated click-logs of 1389 fake accounts and more than a million logs of 45,769 general users in a highly popular web portal. Findings show that the inflated visibility of anti-governmental opinions in the manipulated comment space was incongruent with the overall political tone that general users had spontaneously encountered from the broader media ecosystem beyond the manipulated space. Subsequently, this opinion “climate” incongruence decreased the likelihood of commenting in the manipulated space. The study concludes that false amplification (of the opinions that the manipulators promote) and false diminution (of general users’ political expressions) work in tandem to create a distorted opinion environment.

show abstract

“…This problem is intensified by evidence that there is a large population of susceptible users on social media who are emotionally charged in their public debates on social media [57] and are unable to distinguish between malicious social bots and other human users. Indeed, studies have shown that social bots have an influence on people by swaying them towards certain positions, creating polarization amongst the public, or obfuscating political issues with tangential information [58].…”

Section: Ojcoms-00446-2021mentioning

confidence: 99%

“…These detection techniques include structurebased systems that explore the networks and interactions of known bots, as bots generally interact with other social bots more than with real accounts to generate their fake social communities. Some of these techniques are available freely to the public through such software as BotOrNot, BotOMeter, and others that provide metrics of bot data and bot scores or the likelihood that an account is a bot [58]; The BotOmeter tests the probability of a Twitter account being a social bot based on multiple features. Some of the features that can be used to detect bots include posting and reposting rates, timing of messages, network patterns (followers and friends), sentiment pattern, etc.…”

Section: V2 Social Media Forensicsmentioning

confidence: 99%

“…Another method includes using a learning agent bot to communicate with a random environment and, using multiple techniques listed above, identify other accounts that it interacts with that are most likely social bots [83]. On top of these methods, some researchers use honeypot methods to attract bots and then provide an environment to capture data on their behavioral patterns, as well as researchers who create their own bots to observe patterns [58].…”

Section: V2 Social Media Forensicsmentioning

confidence: 99%

See 1 more Smart Citation

Attribution Across Cyber Attack Types: Network Intrusions and Information Operations

Goel

Nussbaum

2021

IEEE Open J. Commun. Soc.

View full text Add to dashboard Cite

The objective of this paper is to articulate the problem of attribution in cyber warfare incidents, including, surveillance, data theft, espionage, and misinformation campaigns. As the stakes increase, concerted efforts are being made by intelligence and law enforcement agencies to identify the perpetrators with much painstaking effort. Attribution tools and techniques for malicious activities on the internet are still nascent, relying mainly on technical measurements, the provenance of malicious code, and nontechnical assessments of attack and attacker characteristics to link attack activities to individuals or groups. Attribution of attacks is typically done through a burdensome manual process that relies on both technical analysis and ground intelligence. As a result, this cumbersome and laborious process of attribution is primarily reserved for the most egregious cyber attack cases and those conducted against well resourced organizations. Over time, our attribution abilities have improved, however, this improvement is a two-edged sword: as attribution capabilities improve, internet privacy is increasingly diluted. This paper discusses attribution for two vastly different types of attacks that are central to cyber conflict today: network intrusions and social bot-led misinformation campaigns. The paper discusses the state of the art regarding attribution abilities across both types of attack, provides recommendations for improved attribution, and lays out future research directions.

show abstract

Trends in Detection and Characterization of Propaganda Bots

Cited by 10 publications

References 24 publications

TWEETSPIN: Fine-grained Propaganda Detection in Social Media Using Multi-View Representations

TWEETSPIN: Fine-grained Propaganda Detection in Social Media Using Multi-View Representations

Fake thumbs in play: A large-scale exploration of false amplification and false diminution in online news comment spaces

Attribution Across Cyber Attack Types: Network Intrusions and Information Operations

Contact Info

Product

Resources

About