Positive Impact on Equity in Society.
Fair access for historically disadvantaged groups of people to potentially life changing opportunities such as jobs, loans, and educational opportunities is a transformative societal outcome that our AEQUITAS technology aims to facilitate. This project promises to develop methods and tools to help decision makers mitigate both the implicit bias they suffer from as well as expose algorithmic bias inadvertently embedded in automated AI ranking algorithms. Our proposed methods will support the detection of discriminatory practices for tasks such as hiring, facilitating fair algorithmic practices, and life-altering decision making. Broader impact also includes the dissemination of bias mitigation software, empowering others to apply this technology for their decision making.
Integrating Diversity, Outreach and Undergraduate Students.
Prof. Rundensteiner has a long history of involving underrepresented groups in research via CRA mentoring. Well over half of her graduated 36 PhD students over the past 20 years have been female or members of other underrepresented groups. Broadening participation and successful graduation of female and minority students rests on supportive student mentoring provided by the PI. The PI has excelled at mentoring undergraduates in research projects, as also planned for this proposed NSF project.
Integration of Research and Education.
The MS program in interdisciplinary Data Science at WPI started Fall 2014, led by founding director Rundensteiner (PI), and demand has since skyrocketed. Subsequently, the interdisciplinary Data Science PhD, the first in the nation, was approved by WPI Trustees in Fall 2015. This program produces much sought after MS and PhD graduates ready to assume leadership roles in data-enabled STEM careers. PI assembled the Executive Industrial Advisory Board for Data Science from IBM, Dell, MIT LL, MITRE Corp., MathWorks, and other stakeholders. Close collaboration with industry allows the PI to stay abreast with problems faced by industry and share NSF research innovations. The PI and her colleagues rolled out an exemplar qualifying graduate project degree component that embeds all DS graduate students as 3 to 4 person teams into industry-sponsored capstone projects -- a model which WPI is now working on replicating across other majors.
Knowledge developed by our project team will be infused back into our educational agenda using a rich repertoire of mechanisms. Our undergraduate curriculum at WPI will be enriched with compelling fair consensus building projects. As part of the data science undergraduate major being rolled out at WPI, the PI and her colleagues are designing a data science introduction course where students will be exposed to material from this NSF project, in particular on bias embedded in data and risk of its propagation by AI algorithms. CoPI Harrison is designing a new course in Data Visualization at the undergraduate level; with the graduate version already very popular. Innovations in interactive systems derived from this NSF project will be incorporated into future course offerings as projects.