Graph Sparsification for Enhanced Conformal Prediction in Graph Neural Networks

TL;DR

This work introduces SparGCP, which incorporates graph sparsification and a conformal prediction-specific objective into GNN training, and employs a parameterized graph sparsification module to filter out task-irrelevant edges, thereby improving conformal prediction efficiency.

Abstract

Conformal Prediction is a robust framework that ensures reliable coverage across machine learning tasks. Although recent studies have applied conformal prediction to graph neural networks, they have largely emphasized post-hoc prediction set generation. Improving conformal prediction during the training stage remains unaddressed. In this work, we tackle this challenge from a denoising perspective by introducing SparGCP, which incorporates graph sparsification and a conformal prediction-specific objective into GNN training. SparGCP employs a parameterized graph sparsification module to filter out task-irrelevant edges, thereby improving conformal prediction efficiency. Extensive experiments on real-world graph datasets demonstrate that SparGCP outperforms existing methods, reducing prediction set sizes by an average of 32\% and scaling seamlessly to large networks on commodity GPUs.

Figures (7)

• Figure 1: Model Architecture.
• Figure 2: MFG example.
• Figure 4: APS efficiency with varying strength of graph sparsification.
• Figure 5: Varying the weight of CP-based loss.
• Figure 6: Evaluation of different combinations of GNNs (Vanilla GCN and SparGCP (GCN)) and post-hoc CP methods (APS and CF-GNN) in terms of APS efficiency and training time in seconds.