FairSample: Training Fair and Accurate Graph Convolutional Neural Networks Efficiently
This work addresses fairness concerns in GCNs for applications where societal biases exist, representing an incremental improvement by combining existing strategies like edge injection and reinforcement learning with regularization.
The paper tackles the problem of training fair and accurate Graph Convolutional Neural Networks (GCNs) efficiently by addressing biases in graph structure, node attributes, and model parameters, resulting in a framework called FairSample that jointly mitigates these biases.
Fairness in Graph Convolutional Neural Networks (GCNs) becomes a more and more important concern as GCNs are adopted in many crucial applications. Societal biases against sensitive groups may exist in many real world graphs. GCNs trained on those graphs may be vulnerable to being affected by such biases. In this paper, we adopt the well-known fairness notion of demographic parity and tackle the challenge of training fair and accurate GCNs efficiently. We present an in-depth analysis on how graph structure bias, node attribute bias, and model parameters may affect the demographic parity of GCNs. Our insights lead to FairSample, a framework that jointly mitigates the three types of biases. We employ two intuitive strategies to rectify graph structures. First, we inject edges across nodes that are in different sensitive groups but similar in node features. Second, to enhance model fairness and retain model quality, we develop a learnable neighbor sampling policy using reinforcement learning. To address the bias in node features and model parameters, FairSample is complemented by a regularization objective to optimize fairness.