Date of Graduation
5-2025
Document Type
Thesis
Degree Name
Bachelor of Science in Computer Science
Degree Level
Undergraduate
Department
Computer Science and Computer Engineering
Advisor/Mentor
Gauch, Susan
Committee Member
Luu, Khoa
Second Committee Member
Zhang, Lu
Abstract
Bias in academic paper selection remains a consistent issue, even within processes designed to promote fairness, such as double-blind peer review, bias stays persistent. In this paper we investigate demographic bias while particularly focusing on racial bias in the process of selecting academic papers and explore the impact of fairness aware recommender systems on the demographic parity. To build an effective system our focus is on the Special Interest Group on Computer Human Interaction (SIGCHI) a pillar in the community, we develop a neural network-based recommender system that uses real demographic data collected by other systems withing the context of SIGCHI and evaluates fairness from using several methods. Using different parameters such as a fairness value of lambda, our system accepts that there will a tradeoff between utility which is actually measured by the h-index and fairness which is a comparison to the overall paper pool. Through a series of experiments across varying methods and number of papers selected, this paper demonstrates that it is possible to close the gap on bias by improving representation of marginalized groups in paper selection while maintaining appropriate quality of papers.
Keywords
recommender; bias; SIGCHI; PyTorch
Citation
Bergin, Z. (2025). Fairness in Recommender Systems: Balancing Bias in Academic Paper Selection. Electrical Engineering and Computer Science Undergraduate Honors Theses Retrieved from https://scholarworks.uark.edu/elcsuht/6
