analiza aiqbench
This research addresses critical evaluation flaws in multi-task learning by introducing ReLiF, a framework that enforces shared reference tolerances to prevent method-induced bias distortions during auditing. By decoupling fixed-delta verification from training regularization, the authors demonstrate how current fairness metrics can obscure genuine utility-fairness trade-offs across diverse benchmarks like clinical time-series and dense prediction tasks. The work provides essential theoretical conditions for preserving ranking consistency when assessing individual fairness in complex model architectures.
Is Fairness Truly Fair? Towards Reliable Lipschitz Fairness in Multi-Task Learning via Fixed-\texorpdfstring{$\delta$}{delta} Alignment
arXiv:2606.10632v1 Announce Type: new Abstract: Lipschitz-style individual fairness formalizes the idea that semantically similar examples should receive similar predictions, but its evaluation in multi-task learning (MTL) can be confounded by method-induced representation scales. This paper identifies threshold confounding: when the auditing tolerance is derived from each model's own representation distances, different algorithms are compared under different semantic thresholds. A threshold-drift analysis further shows how Bias rankings can change and identifies sufficient conditions for ranking preservation. We propose \textbf{ReLiF}, a reliability-aware framework that separates evaluation-time fixed-$\delta$ auditing from training-time controlled regularization. ReLiF uses a shared reference tolerance for comparable auditing and a violation-rate feedback controller to keep the Lipschitz surrogate active without letting it dominate stochastic training. This work also develops supporting analysis for threshold drift, reference-tolerance selection, and the relationship between the huberized training surrogate and its unsmoothed positive-margin counterpart. Experiments on clinical time-series benchmarks and NYUv2 (NYU Depth V2) dense prediction show that fixed-$\delta$ auditing exposes utility--fairness trade-offs that method-dependent thresholds can obscure. On NYUv2 with a ResNet50 backbone, ReLiF achieves competitive utility while substantially reducing aligned bias under shared fixed thresholds. On clinical benchmarks, ReLiF yields controlled fairness-regularized trade-offs, while fixed-$\delta$ auditing reveals that task-balancing baselines can sometimes achieve lower bias and that genuine utility--fairness trade-offs persist. These results support fixed-$\delta$ auditing as a semantically consistent protocol for evaluating Lipschitz fairness in MTL.
Title: Is Fairness Truly Fair? Towards Reliable Lipschitz Fairness in Multi-Task Learning via Fixed-$\delta$ Alignment
Abstract: Lipschitz-style individual fairness formalizes the idea that semantically similar examples should receive similar predictions, but its evaluation in multi-task learning (MTL) can be confounded by method-induced representation scales. This paper identifies threshold confounding: when the auditing tolerance is derived from each model's own representation distances, different algorithms are compared under different semantic thresholds. A threshold-drift analysis further shows how Bias rankings can change and identifies sufficient conditions for ranking preservation. We propose **ReLiF**, a reliability-aware framework that separates evaluation-time fixed-$\delta$ auditing from training-time controlled regularization. ReLiF uses a shared reference tolerance for comparable auditing and a violation-rate feedback controller to keep the Lipschitz surrogate active without letting it dominate stochastic training. This work also develops supporting analysis for threshold drift, reference-tolerance selection, and the relationship between the huberized training surrogate and its unsmoothed positive-margin counterpart. Experiments on clinical time-series benchmarks and NYUv2 (NYU Depth V2) dense prediction show that fixed-$\delta$ auditing exposes utility--fairness trade-offs that method-dependent thresholds can obscure. On NYUv2 with a ResNet50 backbone, ReLiF achieves competitive utility while substantially reducing aligned bias under shared fixed thresholds. On clinical benchmarks, ReLiF yields controlled fairness-regularized trade-offs, while fixed-$\delta$ auditing reveals that task-balancing baselines can sometimes achieve lower bias and that genuine utility--fairness trade-offs persist. These results support fixed-$\delta$ auditing as a semantically consistent protocol for evaluating Lipschitz fairness in MTL.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2606.10632 [cs.LG]
(or arXiv:2606.10632v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.10632 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.1145/3770855.3817938 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 343 cuvinte
Subiecte conexe
Skydio CEO Adam Bry on why Silicon Valley shouldn’t draw red lines for drone use
The Verge · 15 iun. 2026
Underwriting Superintelligence: Backing Agents you can Sue — Rune Kvist, AIUC
Latent Space · 16 sept. 2026
State of the Field: AI for Epistemics and Coordination
LessWrong · 2 oct. 2026
[AINews] not much happened today
Latent Space · 18 sept. 2026