Human-machine detection of online-based malign information

Marcellino, William; Cox, Kate; Galai, Katerina; Slapakova, Linda; Jaycocks, Amber; Harris, Ruth

doi:10.7249/rra519-1

Cited by 4 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…68 Additionally, the full report explains technical concepts and methods clearly and accessibly and highlights the level of accuracy of the model and interpretability to establish confidence in the results. 69 Another relevant example is past RAND research on using machine learning to address infant mortality. 70 a high-performing model.…”

Section: Best Practices On Communicating the Results Of Machine Learn...mentioning

confidence: 99%

Demographic and Geographic Characteristics of Green Stormwater Infrastructure Investments in Five U.S. Cities: A Machine Learning–Based Analysis

2023

View full text Add to dashboard Cite

■ Although no cities we examined were consistently and deliberately planning based on a co-benefits framework, areas that were more strongly associated with green stormwater infrastructure had higher percentages of Hispanic or Latino residents, lower percentages of White residents, and higher percentages of residents who self-reported coronary health challenges.■ City stormwater planning has predominately been consent decree-driven and publicly funded with a focus on reducing stormwater volumes. It is now, however, shifting toward an effort that is driven by compliance with municipal codes and that includes more private citizens. Ensuring that these more-distributed approaches carefully consider co-benefits could enhance the impact of stormwater management moving forward.

show abstract

Section: Best Practices On Communicating the Results Of Machine Learn...mentioning

confidence: 99%

Demographic and Geographic Characteristics of Green Stormwater Infrastructure Investments in Five U.S. Cities: A Machine Learning–Based Analysis

2023

View full text Add to dashboard Cite

show abstract

“…This line of research can be categorized into (1) detection of their campaigns [12,19,20,21,22,23], (2) characterizing the respective content, goals, and tactics [19,24,25,26,27,28],…”

Section: Introductionmentioning

confidence: 99%

Russian Information Operations May Have Caused Hate Crimes

Alizadeh¹,

Chen²,

Shapiro

2023

Preprint

View full text Add to dashboard Cite

Understanding whether nation-state information operations on social media have real-world effects is an important question. Previous research has shown that Russia targeted American divisions along issues of race, however, it did not have any impact on political attitudes and behaviors of Twitter users. We identify Islam as another issue on which Russian trolls posted inflammatory content and study the causal effect of their tweets about Islam on anti-Muslim hate crimes in the United States. Using events, holidays, and weather conditions in Saint Petersburg, where a key Russian troll organization is based, as an exogenous source of variation in content production, we find that increases in anti-Islam activity by Russian trolls led to increases in anti-Muslin hate crimes in the United States. In our preferred specification, a one standard deviation increase in log(Russian tweets about Islam) predicts a 30% increase from the mean rate of anti-Islam hate crimes for US divisions with average exposure to IRA trolls. Our results have policy implications for combating information operations and safeguarding the public security.

show abstract

“…Also called human-supervised learning, this might, for example, mean inserting examples of threatening language, angry-but-not-threatening language, and neutral language into an ML model and then teaching the model to classify new documents into one of those classes.5 For more on the stance taxonomy, seeRingler, Klebanov, and Kaufer (2018). For prior use of stance in modeling, seeMarcellino et al (2020).…”

mentioning

confidence: 99%

Detecting Conspiracy Theories on Social Media: Improving Machine Learning to Detect and Understand Online Conspiracy Theories

Marcellino¹,

Helmus²,

Kerrigan³

et al. 2021

View full text Add to dashboard Cite

Conspiracy theories circulated online over social media contribute to a shift in public discourse away from facts and analysis and can contribute to direct public harm. Social media platforms are encountering a difficult technical and policy challenge as they try to mitigate harm from online conspiracy theory language. As part of an effort to confront emerging threats and incubate new technology to help create a safer world, Google's Jigsaw unit asked RAND Corporation researchers to conduct a modeling effort to improve machine-learning technology for detecting conspiracy theory language by using linguistic and rhetorical theory to boost performance. We also aimed to synthesize existing research on conspiracy theories with new insight from this improved modeling effort. In this report, we share insights from the effort and offer recommendations to mitigate harm and reduce the effects of conspiracy theories online.The research reported here was completed in January 2021 and underwent security review with the sponsor and the Defense Office of Prepublication and Security Review before public release.

show abstract

Human-machine detection of online-based malign information

Cited by 4 publications

References 10 publications

Demographic and Geographic Characteristics of Green Stormwater Infrastructure Investments in Five U.S. Cities: A Machine Learning–Based Analysis

Demographic and Geographic Characteristics of Green Stormwater Infrastructure Investments in Five U.S. Cities: A Machine Learning–Based Analysis

Russian Information Operations May Have Caused Hate Crimes

Detecting Conspiracy Theories on Social Media: Improving Machine Learning to Detect and Understand Online Conspiracy Theories

Contact Info

Product

Resources

About