An algorithm learns a Mahalanobis metric from triplet queries via spectral initialization and gradient descent in the Bradley-Terry model, with convergence guarantees and transfer of individual fairness from estimated to true metric.
IEEE Transactions on Signal Processing , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Operationalizing Individual Fairness via Gradient Descent and Bradley-Terry Models
An algorithm learns a Mahalanobis metric from triplet queries via spectral initialization and gradient descent in the Bradley-Terry model, with convergence guarantees and transfer of individual fairness from estimated to true metric.