A Unified Approach to Fair Online Learning via Blackwell Approachability - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Advances in Neural Information Processing Systems Year : 2021

A Unified Approach to Fair Online Learning via Blackwell Approachability


We provide a setting and a general approach to fair online learning with stochastic sensitive and non-sensitive contexts. The setting is a repeated game between the Player and Nature, where at each stage both pick actions based on the contexts. Inspired by the notion of unawareness, we assume that the Player can only access the non-sensitive context before making a decision, while we discuss both cases of Nature accessing the sensitive contexts and Nature unaware of the sensitive contexts. Adapting Blackwell's approachability theory to handle the case of an unknown contexts' distribution, we provide a general necessary and sufficient condition for learning objectives to be compatible with some fairness constraints. This condition is instantiated on (group-wise) no-regret and (group-wise) calibration objectives, and on demographic parity as an additional constraint. When the objective is not compatible with the constraint, the provided framework permits to characterise the optimal trade-off between the two.
Fichier principal
Vignette du fichier
Chzhen-Giraud-Stoltz--Online-Fair-Learning-Blackwell.pdf (484.39 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03268279 , version 1 (23-06-2021)
hal-03268279 , version 2 (15-10-2021)



Evgenii Chzhen, Christophe Giraud, Gilles Stoltz. A Unified Approach to Fair Online Learning via Blackwell Approachability. 35th Conference on Neural Information Processing Systems (NeurIPS 2021), 2021, Virtual conference, Australia. ⟨hal-03268279v2⟩
212 View
121 Download



Gmail Mastodon Facebook X LinkedIn More