Towards a Unified Framework for Fair and Stable Graph Representation Learning
Towards a Unified Framework for Fair and Stable Graph Representation Learning
复制标题
DOI：
--
复制
发表时间：
2021-02
期刊：
影响因子：
--
通讯作者：
Chirag Agarwal;Himabindu Lakkaraju;M. Zitnik
中科院分区：
文献类型：
--
作者：
Chirag Agarwal;Himabindu Lakkaraju;M. Zitnik
随着图神经网络（GNN）输出的表示越来越多地用于现实世界的应用，确保这些表示是公平和稳定的变得非常重要。在这项工作中，我们建立了反事实公平性和稳定性之间的关键联系，并利用它提出了一个新的框架，NIFTY（uNIfying Fairness and stabilITY），它可以与任何GNN一起使用，以学习公平和稳定的表示。我们引入了一个新的目标函数，同时考虑了公平性和稳定性，并使用Lipschitz常数开发了一个逐层权重归一化，以增强GNN中的神经消息传递。在这样做的过程中，我们在目标函数和GNN架构中都加强了公平性和稳定性。此外，我们从理论上表明，我们的逐层权重归一化促进了所得到的表示中的反事实公平性和稳定性。我们介绍了三个新的图形数据集，包括刑事司法和金融贷款领域的高风险决策。对上述数据集的广泛实验证明了我们框架的有效性。
As the representations output by Graph Neural Networks (GNNs) are increasingly employed in real-world applications, it becomes important to ensure that these representations are fair and stable. In this work, we establish a key connection between counterfactual fairness and stability and leverage it to propose a novel framework, NIFTY (uNIfying Fairness and stabiliTY), which can be used with any GNN to learn fair and stable representations. We introduce a novel objective function that simultaneously accounts for fairness and stability and develop a layer-wise weight normalization using the Lipschitz constant to enhance neural message passing in GNNs. In doing so, we enforce fairness and stability both in the objective function as well as in the GNN architecture. Further, we show theoretically that our layer-wise weight normalization promotes counterfactual fairness and stability in the resulting representations. We introduce three new graph datasets comprising of high-stakes decisions in criminal justice and financial lending domains. Extensive experimentation with the above datasets demonstrates the efficacy of our framework.