Thomas Davidson
YOU?
Author Swipe
View article: AI reasoning effort predicts human decision time in content moderation
AI reasoning effort predicts human decision time in content moderation Open
Large language models can now generate intermediate reasoning steps before producing answers, improving performance on difficult problems by interactively developing solutions. This study uses a content moderation task to examine parallels…
View article: Auditing multimodal large language models for context-aware content moderation
Auditing multimodal large language models for context-aware content moderation Open
The development of multimodal large language models (MLLMs) offers new possibilities for context-aware content moderation by integrating text, images, and other data. This study uses a series of conjoint experiments to systematically audit…
View article: COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements
COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements Open
Warning: This paper contains content that may be offensive or upsetting. Understanding the harms and offensiveness of statements requires reasoning about the social and situational context in which statements are made. For example, the utt…
View article: Partisan US News Media Representations of Syrian Refugees
Partisan US News Media Representations of Syrian Refugees Open
We investigate how representations of Syrian refugees (2011-2021) differ across US partisan news outlets. We analyze 47,388 articles from the online US media about Syrian refugees to detail differences in reporting between left- and right-…
View article: COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements
COBRA Frames: Contextual Reasoning about Effects and Harms of Offensive Statements Open
Warning: This paper contains content that may be offensive or upsetting. Understanding the harms and offensiveness of statements requires reasoning about the social and situational context in which statements are made. For example, the utt…
View article: Hate Speech Detection and Bias in Supervised Text Classification
Hate Speech Detection and Bias in Supervised Text Classification Open
This chapter introduces hate speech detection, describing the process of developing annotated datasets and training machine learning models and highlighting cutting-edge techniques used to improve these methods. The case of hate speech det…
View article: See it to believe it?
See it to believe it? Open
A common fixture of computer systems research are Visualisation-in-the-Loop Tools: tools that produce complex output data and require a human user to interpret the data visually. However, systems research frequently omits or sidelines deta…
View article: Partisan US News Media Representations of Syrian Refugees
Partisan US News Media Representations of Syrian Refugees Open
We investigate how representations of Syrian refugees (2011-2021) differ across US partisan news outlets. We analyze 47,388 articles from the online US media about Syrian refugees to detail differences in reporting between left- and right-…
View article: Engaging Populism? The Popularity of European Populist Political Parties on Facebook and Twitter, 2010-2020
Engaging Populism? The Popularity of European Populist Political Parties on Facebook and Twitter, 2010-2020 Open
Scholars have argued that populists disproportionately benefit from social media, and there is evidence that they attract more engagement than other politicians and parties in several countries. We systematically evaluate the relationship …
View article: Semi-Supervised Annotation of Portuguese Hate Speech Across Social Media Domains
Semi-Supervised Annotation of Portuguese Hate Speech Across Social Media Domains Open
With the increasing spread of hate speech (HS) on social media, it becomes urgent to develop models that can help detecting it automatically. Typically, such models require large-scale annotated corpora, which are still scarce in languages…
View article: Interface Engineering through Adhesion and Remineralization for Dental Composites using Mineral Binding Peptides
Interface Engineering through Adhesion and Remineralization for Dental Composites using Mineral Binding Peptides Open
Nearly 90% of adults in the US will develop dental caries needing treatment with dental restorations within their lifetimes. An increasing number of restorations are done using composite rather than amalgams. Dental composites have some be…
View article: Aggregate-Driven Trace Visualizations for Performance Debugging
Aggregate-Driven Trace Visualizations for Performance Debugging Open
Performance issues in cloud systems are hard to debug. Distributed tracing is a widely adopted approach that gives engineers visibility into cloud systems. Existing trace analysis approaches focus on debugging single request correctness is…
View article: Britain First and the UK Independence Party: Social Media and Movement-Party Dynamics
Britain First and the UK Independence Party: Social Media and Movement-Party Dynamics Open
Social movement scholars have recently turned their attention to the interactions between political parties and social movements, but little is known about how social media have impacted these relationships, despite widespread adoption of …
View article: Variation in Caste Homophily Across Villages and Contexts in Rural India
Variation in Caste Homophily Across Villages and Contexts in Rural India Open
This paper studies on caste homophily across seventy-five villages in rural Karnataka, India. Caste is conceptualized as a system of ethnic categories that map onto positions within the social structure of each village. Despite being a sal…
View article: Black Box Models and Sociological Explanations: Predicting GPA Using Neural Networks
Black Box Models and Sociological Explanations: Predicting GPA Using Neural Networks Open
The Fragile Families Challenge provided an opportunity to empirically assess the applicability of black box machine learning models to sociological questions and the extent to which interpretable explanations can be extracted from these mo…