Seven Months with the Devils: A Long-Term Study of Content Polluters on Twitter

By Kyumin Lee, Brian David Eoff, James Caverlee. International Conference on Web and Social Media, 2011

556 citations recorded by the source corpus.

The rise in popularity of social networking sites such as Twitter and Facebook has been paralleled by the rise of unwanted, disruptive entities on these networks- — including spammers, malware disseminators, and other content polluters. Inspired by sociologists working to ensure the success of commons and criminologists focused on deterring vandalism and preventing crime, we present the first long-term study of social honeypots for tempting, profiling, and filtering content polluters in social media. Concretely, we report on our experiences via a seven-month deployment of 60 honeypots on Twitter that resulted in the harvesting of 36,000 candidate content polluters. As part of our study, we (1) examine the harvested Twitter users, including an analysis of link payloads, user behavior over time, and followers/following network dynamics and (2) evaluate a wide range of features to investigate the effectiveness of automatic content polluter identification.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/1dd5355e62b9fc37a355e135d5909ed28128d653

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

Citations in the corpus

BotOrNot: A System to Evaluate Social Bots (2016, methodology): “As a proof of concept, we used the list of social bots identified by Caverlee’s team [3].”
Cultural Dimensions in Twitter: Time, Individualism and Power (2013, methodology): “The problem is that Hall did not provide any country scores we could use for this study but “Levine’s Pace of Life research has been indirectly linked to the observations of Hall (1983) to suggest that polychronicity and Pace of Life are negatively related” (Conte and Rizzuto 1999), and that insight”
Misinformation in Social Media: Definition, Manipulation, and Detection (2019, methodology): “In [21], the authors use a honeynet containing 60 honeypots to tempt bots to follow them.”
Uncovering social network sybils in the wild (2011, background): “Two studies leverage honeypot accounts on MySpace and Twitter, respectively, to trap spammers who attempt to friend them [Webb et al. 2008; Lee et al. 2011].”
Scalable and Generalizable Social Bot Detection through Data Selection (2019, methodology): “The caverlee dataset consists of bot accounts lured by honeypot accounts and veriﬁed human accounts (Lee, Eoff, and Caverlee 2011).”
The future of social media in marketing (2019, background): “Indeed, social bots (computer algorithms that automatically produce content and interact with social media users; Ferrara et al. 2016) have inhabited social media platforms for the last decade (Lee et al. 2011), and have become increasingly pervasive.”
Tampering with Twitter’s Sample API (2018, background): “We face a variety of biases in social media data, such as targeted user collaboration to skew contents of a site [55], changing statistics on site like trends and follower counts in an attempt to skew the perception of real users on the site [56].”
Deep Neural Networks for Bot Detection (2018, methodology): “rs used clustering techniques to identify large-scale behavioral anomalies [56], while other authors adopted supervised learning to analyze all accounts of some platform and separate bots from humans [7, 17, 37, 60]. Although these approaches can be useful, for example to detect large-scale bot infi”
The DARPA Twitter Bot Challenge (2016, unknown): “However, with the exception of [3], no past work has looked specifically at identifying influence bots on a specific topic.”
Understanding and combating link farming in the twitter social network (2012, background): “Also, a number of recent studies report an increasing number of spammers inﬁltrating the Twitter network [20].”
Finding and Characterizing Information Warfare Campaigns (2020, methodology): “The Varol data set was founded on the original 2011 Caverlee [150] Honey Pot data, but was supplemented with manual annotations (we leveraged only the manually annotated data).”
Online Human-Bot Interactions: Detection, Estimation, and Characterization (2017, methodology): “Lee et al. used a honeypot approach to collect the largest sample of bot accounts available to date (Lee, Eoff, and Caverlee 2011).”
The rise of social bots (2014, background): “Social bots have been circulating on social media platforms for a few years [Lee et al. 2011; Boshmaf et al. 2011].”
Arming the public with artificial intelligence to counter social bots (2019, methodology): “In 2011, Lee et al. (2011) identiﬁed thousands of social bots.”
Data ex Machina: Introduction to Big Data (2017, background): “Robots and organizations engage on these platforms for a variety of prosocial and antisocial purposes, making them difficult to detect and theorize (Lee et al. 2011, Ferrara 2015).”
Explainable Artificial Intelligence Applications in Cyber Security: State-of-the-Art in Research (2022, methodology): “The obtained detection rate 1153 of Honeypot is 97.2% and the NSL-KDD dataset is 95.8% 1154 which was superior and LIME and SHAP were deployed to 1155 explain the detection decisions.”
The Paradigm-Shift of Social Spambots: Evidence, Theories, and Tools for the Arms Race (2017, methodology): “Specifically, it employs off-theshelf supervised learning algorithms trained with examples of both humans and bots behaviors, based on the Texas A&M dataset [25] with 15,000 examples of each class and millions of tweets.”