Balanced Neighborhoods for Multi-sided Fairness in Recommendation
Association for Computing Machinery (ACM) via YouTube
Overview
Explore a conference talk that delves into the concept of balanced neighborhoods for multi-sided fairness in recommendation systems. Learn about the unique aspects of this research, the filter bubble phenomenon, and the necessity of recommendations. Examine various considerations, including the Kiva example and bias in data. Gain insights into balanced neighborhoods, sensitivity analysis, and draw conclusions on improving fairness in recommendation algorithms. This 19-minute presentation by Robin Burke from DePaul University, delivered at FAT* 2018, offers valuable perspectives on addressing fairness issues in modern recommendation systems.
Syllabus
Introduction
What is different about this work
The filter bubble
Do we really need recommendation
Other considerations
Kiva example
Bias in data
Balanced neighborhoods
Sensitivity
Conclusion
Taught by
ACM FAccT Conference