Towards Fair Graph Prompting: A Dual-Prompt Mechanism for Mitigating Attribute and Structural Bias
Researchers have introduced Adaptive Dual Prompting (ADPrompt), a new framework designed to make Graph Neural Networks (GNNs) fairer. Existing methods often overlook biases present in graph data, leading to unfair outcomes in downstream tasks. ADPrompt addresses this by using personalized attribute prompts to reduce sensitive information and structural prompts to control information flow, thereby mitigating both attribute and structural biases. Experiments show ADPrompt outperforms existing methods on node classification tasks. AI
IMPACT Enhances fairness in graph-based AI models, potentially leading to more equitable outcomes in applications using graph data.