The work “Bounded-Abstention Pairwise Learning to Rank” with Antonio Ferrara, Francesco Bonchi and Salvatore Ruggieri has been accepted at KDD 2026!
This paper studies how to extend abstention mechanisms to the pairwise learning to rank with ties. We show the optimal strategy for abstention relies on thresholding the conditional pairwise risk and then propose practical algorithms to pass from a ranker to a pairwise abstention strategy. We finally showcase the effectiveness of our approach in an extensive experimental evaluation.
A final thanks to all the amazing co-authors involved in this project.
See you in Jeju 🌴⛱️🌊!