Adaptive Dual Prompting: Hierarchical Debiasing for Fairness-aware Graph Neural Networks
Yuhan Yang, Xingbo Fu, Jundong Li
TL;DR
This work addresses fairness gaps that arise when adapting pre-trained Graph Neural Networks via prompting. It introduces Adaptive Dual Prompting (ADPrompt), a framework combining Adaptive Feature Rectification (AFR) at the input layer and Adaptive Message Calibration (AMC) at intermediate layers to suppress attribute and structure bias, respectively, while jointly optimizing with an adversarial objective. Theoretical analysis decomposes the fairness bound into initial bias and bias amplification, showing AFR reduces the former and AMC mitigates the latter, yielding a tighter bound on generalized statistical parity. Extensive experiments across four datasets and four pre-training strategies demonstrate that ADPrompt improves node classification performance while reducing fairness gaps ($\Delta\mathrm{EO}$, $\Delta\mathrm{SP}$) compared to seven baselines. The approach offers a scalable, information-theoretically grounded method for fair graph representation learning with practical applicability to downstream tasks.
Abstract
In recent years, pre-training Graph Neural Networks (GNNs) through self-supervised learning on unlabeled graph data has emerged as a widely adopted paradigm in graph learning. Although the paradigm is effective for pre-training powerful GNN models, the objective gap often exists between pre-training and downstream tasks. To bridge this gap, graph prompting adapts pre-trained GNN models to specific downstream tasks with extra learnable prompts while keeping the pre-trained GNN models frozen. As recent graph prompting methods largely focus on enhancing model utility on downstream tasks, they often overlook fairness concerns when designing prompts for adaptation. In fact, pre-trained GNN models will produce discriminative node representations across demographic subgroups, as downstream graph data inherently contains biases in both node attributes and graph structures. To address this issue, we propose an Adaptive Dual Prompting (ADPrompt) framework that enhances fairness for adapting pre-trained GNN models to downstream tasks. To mitigate attribute bias, we design an Adaptive Feature Rectification module that learns customized attribute prompts to suppress sensitive information at the input layer, reducing bias at the source. Afterward, we propose an Adaptive Message Calibration module that generates structure prompts at each layer, which adjust the message from neighboring nodes to enable dynamic and soft calibration of the information flow. Finally, ADPrompt jointly optimizes the two prompting modules to adapt the pre-trained GNN while enhancing fairness. We conduct extensive experiments on four datasets with four pre-training strategies to evaluate the performance of ADPrompt. The results demonstrate that our proposed ADPrompt outperforms seven baseline methods on node classification tasks.