{"id":"arxiv/2605.23145","url":"https://arxiv.org/abs/2605.23145","tags":["stat.ML","cs.LG","math.ST","stat.ME"],"score":0,"title":"Operationalizing Individual Fairness via Gradient Descent and Bradley-Terry Models","domain":"stat.ML","source":"arxiv","authors":["Conlan Olson","Linjun Zhang","Zhun Deng","Pragya Sur"],"reviews":0,"subject":"stat.ML","summary":"Individual fairness, the notion that \"similar individuals should be treated similarly,\" provides a strong and flexible fairness guarantee for algorithmic decision makers. However, a barrier to implementing individual fairness in practice is the difficulty of learning the similarity metric over individuals. In this work, we present an algorithm for learning a Mahalanobis similarity metric from triplet queries of the form \"is individual $i$ more similar to individual $j$ or $k$?\" We work in the standard Bradley-Terry model for pairwise comparisons. Our algorithm consists of a spectral…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-05-22","reactions":{}}