Pith. sign in

Fairness-Aware Meta-Learning via Nash Bargaining

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

To address issues of group-level fairness in machine learning, it is natural to adjust model parameters based on specific fairness objectives over a sensitive-attributed validation set. Such an adjustment procedure can be cast within a meta-learning framework. However, naive integration of fairness goals via meta-learning can cause hypergradient conflicts for subgroups, resulting in unstable convergence and compromising model performance and fairness. To navigate this issue, we frame the resolution of hypergradient conflicts as a multi-player cooperative bargaining game. We introduce a two-stage meta-learning framework in which the first stage involves the use of a Nash Bargaining Solution (NBS) to resolve hypergradient conflicts and steer the model toward the Pareto front, and the second stage optimizes with respect to specific fairness goals. Our method is supported by theoretical results, notably a proof of the NBS for gradient aggregation free from linear independence assumptions, a proof of Pareto improvement, and a proof of monotonic improvement in validation loss. We also show empirical effects across various fairness objectives in six key fairness datasets and two image classification tasks.

fields

cs.CV 1

years

2024 1

verdicts

CONDITIONAL 1

representative citing papers

MUNBa: Machine Unlearning via Nash Bargaining

cs.CV · 2024-11-23 · conditional · novelty 5.0

MUNBa is a machine unlearning method that uses Nash bargaining to balance forgetting and preservation gradients, improving unlearning quality, generalization, and robustness in image classification and generation.

citing papers explorer

Showing 1 of 1 citing paper.

  • MUNBa: Machine Unlearning via Nash Bargaining cs.CV · 2024-11-23 · conditional · none · ref 78 · internal anchor

    MUNBa is a machine unlearning method that uses Nash bargaining to balance forgetting and preservation gradients, improving unlearning quality, generalization, and robustness in image classification and generation.