Machine Learning for Socio-technical Systems Lab
A university research lab at the University of Rhode Island directed by Dr. Sarah M Brown, studying how machine learning interacts with complex socio-technical systems, with a focus on fairness of automated decision-making and AI safety evaluation.
A university research lab at the University of Rhode Island directed by Dr. Sarah M Brown, studying how machine learning interacts with complex socio-technical systems, with a focus on fairness of automated decision-making and AI safety evaluation.
People
Updated 05/18/26Director
Undergraduate researcher
PhD student
PhD student
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The Machine Learning for Socio-technical Systems Lab (ML4STS) is a research lab in the Department of Computer Science and Statistics at the University of Rhode Island, directed by Dr. Sarah M Brown. The lab was established in 2020 when Dr. Brown joined URI as an Assistant Professor. The lab's mission centers on understanding how machine learning interacts with the complex socio-technical systems into which algorithms are deployed. The research program has three main thrusts: applying data science using ML models to understand the world in collaboration with domain scientists, studying and evaluating ML systems directly, and building tools to help data scientists employ best practices. The majority of the lab's work focuses on fairness of automated decision-making systems. Key active projects include developing a benchmark for evaluating fair data-driven decision-making with large language models (funded by the Survival and Flourishing Fund), model-based fairness intervention assessment, task-level fairness evaluation, and studying perceptions of AI fairness in collaboration with Brown University. The lab also works on simulation-based AI safety curricula for students. Dr. Brown's background includes a PhD in Electrical Engineering from Northeastern University, a postdoctoral fellowship at UC Berkeley under Professor Michael Jordan, and a Data Science Initiative Postdoc at Brown University. She is the treasurer emeritus of Women in Machine Learning, Inc., and an NSF Graduate Research Fellowship recipient. The lab received an NSF CAREER Award in 2025 worth $600,000 over five years for the project "Realizing Sociotechnical Machine Learning through Modeling, Explanations, and Reflections." Other funding includes a 2021 IBM Global University Program Academic Award and a grant from the Survival and Flourishing Fund's Fairness Track. The lab team currently includes two PhD students and approximately eight undergraduate researchers, along with a growing list of alumni.
Theory of Change
Updated 05/18/26The ML4STS Lab works to reduce harms from automated decision-making systems by developing practical tools, benchmarks, and frameworks that help data scientists identify and prevent unintended systemic errors before deployment. By studying fairness at the task level, building evaluation benchmarks for LLM-based decision-making, and creating interventions informed by both technical and social science perspectives, the lab aims to ensure ML systems do not reinforce patterns of discrimination. The lab also develops AI safety curricula for students, helping train the next generation of technologists to anticipate potential societal impacts of AI systems during the design phase rather than after deployment.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26Past project examining how notions of fairness generalize, tolerate distribution shift, and propagate through interacting systems, currently inactive and seeking a student.
Project developing a benchmark to test how LLMs and LLM-based agents make non-discriminatory, fair decisions from data, including direct decision-making, assisting fair ML tasks, and agentic model training.
Project that uses bias models to evaluate how different fair machine learning interventions perform, with the goal of giving data scientists more actionable guidance when choosing fairness techniques.
Collaborative project with C. Malik Boykin’s lab at Brown University studying people’s preferred definitions of algorithmic fairness and how social and algorithmic factors shape those preferences, including methods to interpolate between fairness definitions.
Discussion
No comments yet. Be the first to share your thoughts.