Funding provided to projects that seek to solve social media misinformation, vulnerable community infrastructure, and racial injustice
The SoDa-Social Data Science Center has just finished awarding its inaugural 2022 SoDa seed grant. Five projects received these awards. Winning projects are cross-disciplinary, affecting large groups or communities of people, and stimulate the advancement of new ideas that can build the university’s reputation as an expert in social data science.
About the projects
Giovanni Ciampaglia: Creating a Trustworthy Social Media Platform
Giovanni Ciampaglia will use the seed grant to develop Rockwell, a simulated social media environment, to run field-like experiments on algorithms that recommend social media content, which will be programmed only to supply trustworthy information to social media users regardless of their preferences.
Charles Harry: Examining Vulnerability of Community Critical Infrastructure
Local government-controlled critical infrastructure relies on computer systems and is increasingly vulnerable to cyber-attacks that can disrupt operations. This project aims to examine the relationship between cyber vulnerabilities in some critical infrastructure sectors in Maryland counties with available demographic factors to identify potential cyber security disparities across different communities in the state. The results of this analysis will serve as an initial examination of the distribution of the potential of strategic vulnerability to community critical infrastructure. They can be used to drive the efficient allocation of state resources to improve the security of vulnerable populations.
Rob Wells (JOUR): Racial Terror Lynching and the Press, 1789-1963
This research will help identify structural racism in newspaper reporting by examining racial terror lynching news coverage stretching from 1789 through 1963. This project aims to discover the main stories in white-owned newspapers involving non-court-related killings of Black citizens during this time and then identify whether these stories persist in modern news coverage.
Tracy Sweet (EDUC): How Can Data Science Be Used For Racial Equity?
Current methods in data science may not be effective for all racial groups and could be harmful for some groups. This research aims to identify effective research methods and analyses that can be used to quantify racial disparities in data science analyses and measurement models.
Philip Resnik (ARHU): Effective Few-Shot Learning for Constructs in Psychological and Social Science
In the social sciences, constructs — abstract categories like empathy, misinformation, or benefits of social interaction — are carefully thought through; they are the vocabulary over which theories are defined. This stands in striking contrast to most computational research in AI and natural language processing, where theory tends to be secondary. This project is aiming to bridge that gap by advancing the state of the art in a recent AI approach called few-shot learning, bringing together cutting-edge machine learning with subject matter expertise. Substantively the research focuses on two societally high- impact problems: improved understanding of disparate impacts of COVID-19 across demographic groups by using survey methods with open-ended text responses, and identifying constructs associated with suicidal crisis based on people’s language use.