Robust Graph Fine-Tuning with Adversarial Graph Prompting

TL;DR

This paper introduces a novel adversarial graph prompting framework that enhances the robustness of pre-trained GNN models against noise and attacks through a min-max optimization approach.

Contribution

It proposes the first adversarial graph prompting method with a min-max formulation and an alternating optimization scheme for robust graph fine-tuning.

Findings

AGP improves robustness against graph noise and attacks

Experimental results outperform state-of-the-art methods

Theoretical analysis confirms effectiveness against topology and feature noise

Abstract

Parameter-Efficient Fine-Tuning (PEFT) method has emerged as a dominant paradigm for adapting pre-trained GNN models to downstream tasks. However, existing PEFT methods usually exhibit significant vulnerability to various noise and attacks on graph topology and node attributes/features. To address this issue, for the first time, we propose integrating adversarial learning into graph prompting and develop a novel Adversarial Graph Prompting (AGP) framework to achieve robust graph fine-tuning. Our AGP has two key aspects. First, we propose the general problem formulation of AGP as a min-max optimization problem and develop an alternating optimization scheme to solve it. For inner maximization, we propose Joint Projected Gradient Descent (JointPGD) algorithm to generate strong adversarial noise. For outer minimization, we employ a simple yet effective module to learn the optimal node prompts to counteract the adversarial noise. Second, we demonstrate that the proposed AGP can theoretically address both graph topology and node noise. This confirms the versatility and robustness of our AGP fine-tuning method across various graph noise. Note that, the proposed AGP is a general method that can be integrated with various pre-trained GNN models to enhance their robustness on the downstream tasks. Extensive experiments on multiple benchmark tasks validate the robustness and effectiveness of AGP method compared to state-of-the-art methods.