Revisiting fair-PAC learning and the axioms of cardinal welfare
C Cousins - International Conference on Artificial …, 2023 - proceedings.mlr.press
Cardinal objectives serve as intuitive targets in fair machine learning by summarizing utility
(welfare) or disutility (malfare) $ u $ over $ g $ groups. Under standard axioms, all welfare …
(welfare) or disutility (malfare) $ u $ over $ g $ groups. Under standard axioms, all welfare …
To Pool or Not To Pool: Analyzing the Regularizing Effects of Group-Fair Training on Shared Models
In fair machine learning, one source of performance disparities between groups is overfitting
to groups with relatively few training samples. We derive group-specific bounds on the …
to groups with relatively few training samples. We derive group-specific bounds on the …
Algorithms and analysis for optimizing robust objectives in fair machine learning
C Cousins - arXiv preprint arXiv:2404.06703, 2024 - arxiv.org
The original position or veil of ignorance argument of John Rawls, perhaps the most famous
argument for egalitarianism, states that our concept of fairness, justice, or welfare should be …
argument for egalitarianism, states that our concept of fairness, justice, or welfare should be …
[PDF][PDF] Fair E3: Efficient welfare-centric fair reinforcement learning
As the negative societal consequences of machine learning systems run amok have
become increasingly apparent, fair machine learning methods have seen increased …
become increasingly apparent, fair machine learning methods have seen increased …
[PDF][PDF] Learning Properties in Simulation-Based Games
C Cousins, B Mishra, E Areyan Viqueira… - Proceedings of the …, 2023 - southampton.ac.uk
In recent years, empirical game-theoretic analysis (EGTA) has emerged as a powerful tool
by which to analyze multiagent systems [2, 22, 23, 28], particularly when only a simulator of …
by which to analyze multiagent systems [2, 22, 23, 28], particularly when only a simulator of …
Dividing Good and Great Items Among Agents with Bivalued Submodular Valuations
We study the problem of fairly allocating a set of indivisible goods among agents with
bivalued submodular valuations—each good provides a marginal gain of either a or b (a< b) …
bivalued submodular valuations—each good provides a marginal gain of either a or b (a< b) …
[PDF][PDF] Decentering imputation: fair learning at the margins of demographics
E Dong, C Cousins - Queer in AI Workshop@ ICML, 2022 - cs.brown.edu
Defining demographic categories for fair learning frequently requires both private
information collection at the individual level and expert judgment in the construction of …
information collection at the individual level and expert judgment in the construction of …
Computational and Data Requirements for Learning Generic Properties of Simulation-Based Games
C Cousins, B Mishra, EA Viqueira… - arXiv preprint arXiv …, 2022 - arxiv.org
Empirical game-theoretic analysis (EGTA) is primarily focused on learning the equilibria of
simulation-based games. Recent approaches have tackled this problem by learning a …
simulation-based games. Recent approaches have tackled this problem by learning a …
Dividing Good and Better Items Among Agents with Bivalued Submodular Valuations
We study the problem of fairly allocating a set of indivisible goods among agents with {\em
bivalued submodular valuations}--each good provides a marginal gain of either $ a $ or $ b …
bivalued submodular valuations}--each good provides a marginal gain of either $ a $ or $ b …
[PDF][PDF] On Welfare-Centric Fair Reinforcement Learning
We propose a welfare-centric fair reinforcement-learning setting, in which an agent enjoys
vector-valued reward from a set of beneficiaries. Given a welfare function W (·), the task is to …
vector-valued reward from a set of beneficiaries. Given a welfare function W (·), the task is to …