Skip to main content
QUICK REVIEW

[论文解读] FAIRER: Fairness as Decision Rationale Alignment

Tianlin Li, Qing Guo|arXiv (Cornell University)|Jun 27, 2023
Ethics and Social Impacts of AI被引用 5
一句话总结

该论文提出 FAIRER,一种新颖的公平性框架,将公平性建模为决策理由对齐,确保在 DNN 推理过程中跨子组的一致神经元级影响。通过引入梯度引导的均等对齐以正则化中间神经元响应,FAIRER 在仅造成最小准确率下降的情况下显著提升了公平性,在多个数据集上优于现有方法,其核心在于显式对齐决策理由而非仅最终预测结果。

ABSTRACT

Deep neural networks (DNNs) have made significant progress, but often suffer from fairness issues, as deep models typically show distinct accuracy differences among certain subgroups (e.g., males and females). Existing research addresses this critical issue by employing fairness-aware loss functions to constrain the last-layer outputs and directly regularize DNNs. Although the fairness of DNNs is improved, it is unclear how the trained network makes a fair prediction, which limits future fairness improvements. In this paper, we investigate fairness from the perspective of decision rationale and define the parameter parity score to characterize the fair decision process of networks by analyzing neuron influence in various subgroups. Extensive empirical studies show that the unfair issue could arise from the unaligned decision rationales of subgroups. Existing fairness regularization terms fail to achieve decision rationale alignment because they only constrain last-layer outputs while ignoring intermediate neuron alignment. To address the issue, we formulate the fairness as a new task, i.e., decision rationale alignment that requires DNNs' neurons to have consistent responses on subgroups at both intermediate processes and the final prediction. To make this idea practical during optimization, we relax the naive objective function and propose gradient-guided parity alignment, which encourages gradient-weighted consistency of neurons across subgroups. Extensive experiments on a variety of datasets show that our method can significantly enhance fairness while sustaining a high level of accuracy and outperforming other approaches by a wide margin.

研究动机与目标

  • 为解决公平性正则化 DNN 中的可解释性不足问题,即公平性提升但缺乏对决策机制的理解。
  • 探究公平性是否本质上与不同子组之间决策理由的一致性——即神经元级影响——密切相关。
  • 提出一种新的公平性范式,对齐中间决策理由而非仅最终预测,以提升公平性的鲁棒性。
  • 开发一种实用的训练时间正则化方法,实现实时决策理由对齐,且计算成本可控。

提出的方法

  • 引入参数均等分数,通过测量移除中间神经元时损失的变化,量化不同子组间决策理由的偏移程度。
  • 提出一种新的公平性目标——决策理由对齐(DRAlign),要求在中间层和最终层均实现子组间一致的神经元影响。
  • 提出梯度引导的均等对齐,作为朴素 DRAlign 目标的一种松弛形式,利用样本级梯度而非全数据集计算来估计神经元影响。
  • 通过使用二阶梯度将正则化项整合到训练过程中,以鼓励子组间一致的神经元响应。
  • 采用基于样本的估计策略,使训练期间的计算可处理,从而与标准深度学习优化方法兼容。

实验结果

研究问题

  • RQ1能否通过在不同子组之间对齐神经元的决策理由来提升 DNN 的公平性?
  • RQ2为何现有公平性正则化方法在提升最终预测公平性的同时,仍无法完全消除偏差?
  • RQ3基于中间神经元影响的训练时间正则化,是否能比仅约束最后一层的约束带来更鲁棒的公平性?
  • RQ4如何在 DNN 训练过程中以可接受的计算成本,实际实现决策理由对齐?

主要发现

  • 参数均等分数与模型公平性高度相关,表明子组间决策理由不一致是不公平的根本原因。
  • 现有公平性正则化方法仅部分对齐决策理由,因其仅约束最终层输出,而未控制中间神经元的行为。
  • FAIRER 在三个公开数据集上显著优于最先进方法,同时保持或略微提升准确率。
  • 梯度引导的均等对齐方法实现了有效且可扩展的训练时间优化,无需在每一步都重新评估完整数据集。

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。