FairEGM: Fair Link Prediction and Recommendation via Emulated Graph Modification
FairEGM: Fair Link Prediction and Recommendation via Emulated Graph Modification
复制标题
DOI：
10.1145/3551624.3555287
复制
发表时间：
2022-01
期刊：
影响因子：
--
通讯作者：
Sean Current;Yuntian He;Saket Gurukar;Srinivas Parthasarathy
中科院分区：
文献类型：
--
作者：
Sean Current;Yuntian He;Saket Gurukar;Srinivas Parthasarathy
随着机器学习在各个领域的应用越来越广泛，研究人员和机器学习工程师必须考虑数据中可能由模型永久化的固有偏见。最近，许多研究表明，如果输入图有偏见，这种偏见也会被吸收到图神经网络（GNN）模型中，这可能会对服务不足和代表性不足的社区不利。在这项工作中，我们的目标是通过联合优化两个不同的损失函数来减轻GNN学习到的偏差：一个用于链接预测任务，另一个用于人口统计平价任务。我们进一步实现了三种不同的技术启发图修改方法：全局公平优化（GFO），约束公平优化（CFO）和公平边缘加权（FEW）模型。这些技术模拟了GNN中改变底层图结构的效果，并提供了比更集成的神经网络方法更高程度的可解释性。我们提出的模型模拟了对输入图的微观或宏观编辑，同时训练GNN并学习在链接推荐的背景下准确和公平的节点嵌入。我们在四个真实的世界数据集上证明了我们的方法的有效性，并表明我们可以通过几个因素提高推荐公平性，而链接预测精度的成本可以忽略不计。
As machine learning becomes more widely adopted across domains, it is critical that researchers and ML engineers think about the inherent biases in the data that may be perpetuated by the model. Recently, many studies have shown that such biases are also imbibed in Graph Neural Network (GNN) models if the input graph is biased, potentially to the disadvantage of underserved and underrepresented communities. In this work, we aim to mitigate the bias learned by GNNs by jointly optimizing two different loss functions: one for the task of link prediction and one for the task of demographic parity. We further implement three different techniques inspired by graph modification approaches: the Global Fairness Optimization (GFO), Constrained Fairness Optimization (CFO), and Fair Edge Weighting (FEW) models. These techniques mimic the effects of changing underlying graph structures within the GNN and offer a greater degree of interpretability over more integrated neural network methods. Our proposed models emulate microscopic or macroscopic edits to the input graph while training GNNs and learn node embeddings that are both accurate and fair under the context of link recommendations. We demonstrate the effectiveness of our approach on four real world datasets and show that we can improve the recommendation fairness by several factors at negligible cost to link prediction accuracy.