Watch a research presentation from the 2022 Symposium on Foundations of Responsible Computing where Rojin Rezvan from the University of Texas at Austin explores the design of fair-sponsored search auctions for multiple advertising slots. Learn about extending previous single-slot auction models to handle multiple slots while maintaining fairness and quality tradeoffs in sponsored search settings. Discover how multiplicatively separable click-through rates affect auction design when considering both advertiser-specific and slot-specific components. Examine cases where similar users have comparable click-through rates versus scenarios with varying advertiser preferences, and understand the implementation of preference-based fairness guarantees. Explore a new computationally efficient algorithm for calculating auction payments that builds upon and resolves open questions from previous research in this field.
Overview
Syllabus
Rojin Rezvan | Individually-Fair Auctions for Multi-Slot Sponsored Search
Taught by
Harvard CMSA