Our Paper on Fairness APIs Accepted at ACM TOSEM! 🎉
Thrilled to announce that our paper, “Applications and Challenges of Fairness APIs in Machine Learning Software,” has been accepted for publication in the prestigious ACM Transactions on Software Engineering and Methodology (TOSEM), a top journal in the software engineering field.
This work began as part of my Master’s thesis at the University of Calgary, and I’m proud to see it reach this milestone. 🎓
Paper Summary
We investigated how developers use open-source tools (APIs) to build fairer AI systems by analyzing 204 GitHub projects and over 4,200 developer discussions.
Key takeaways:
- Fairness APIs are being used in highly sensitive areas like healthcare, legal, and business decision-making systems.
- Developers often focus on fairness for groups but tend to overlook potential bias against individuals or subgroups.
- Many developers are still learning and face challenges in understanding fairness concepts, troubleshooting technical issues, and finding the right resources.
Our goal is that these findings can help researchers, educators, and practitioners build more responsible and equitable AI for everyone.
This work would not have been possible without my incredible collaborators. A huge thank you to my MS supervisor, Dr. Gias Uddin, for his invaluable guidance and for taking the lead in wrapping up the final phases of revisions as I became busy with my current professional responsibilities. My heartfelt thanks also go to Dr. Shaiful Alam Chowdhury for working so hard and actively supporting us in shaping this paper.