Similarity-Navigated Conformal Prediction for Graph Neural Networks

TL;DR

This paper introduces SNAPS, a novel similarity-based conformal prediction method for graph neural networks that improves prediction set efficiency and singleton accuracy while guaranteeing coverage.

Contribution

We propose SNAPS, a new algorithm that leverages feature and structural similarity to enhance conformal prediction for GNNs with theoretical coverage guarantees.

Findings

SNAPS produces more compact prediction sets.

SNAPS increases singleton hit ratio.

SNAPS maintains valid coverage guarantees.

Abstract

Graph Neural Networks have achieved remarkable accuracy in semi-supervised node classification tasks. However, these results lack reliable uncertainty estimates. Conformal prediction methods provide a theoretical guarantee for node classification tasks, ensuring that the conformal prediction set contains the ground-truth label with a desired probability (e.g., 95%). In this paper, we empirically show that for each node, aggregating the non-conformity scores of nodes with the same label can improve the efficiency of conformal prediction sets while maintaining valid marginal coverage. This observation motivates us to propose a novel algorithm named Similarity-Navigated Adaptive Prediction Sets (SNAPS), which aggregates the non-conformity scores based on feature similarity and structural neighborhood. The key idea behind SNAPS is that nodes with high feature similarity or direct connections tend to have the same label. By incorporating adaptive similar nodes information, SNAPS can generate compact prediction sets and increase the singleton hit ratio (correct prediction sets of size one). Moreover, we theoretically provide a finite-sample coverage guarantee of SNAPS. Extensive experiments demonstrate the superiority of SNAPS, improving the efficiency of prediction sets and singleton hit ratio while maintaining valid coverage.