Learning and Editing Universal Graph Prompt Tuning via Reinforcement Learning

TL;DR

This work reinforces the theoretical necessity of adding prompts to all nodes for universal graph prompt tuning and introduces LEAP, a learning-and-editing framework that preserves universality while enabling node-wise prompt refinement through actor-critic reinforcement learning. LEAP combines a basic universal graph prompt with RL-driven editing to achieve superior performance across graph- and node-level tasks under diverse pre-training strategies, in both full-shot and few-shot settings. The paper provides extensive empirical evaluations, ablations, and hyper-parameter analyses, demonstrating LEAP's robustness and efficiency, and argues that it establishes an ideal paradigm for universal graph prompt tuning. Overall, LEAP offers a principled, scalable approach to learn high-quality prompts that maintain theoretical guarantees while delivering practical gains.

Abstract

Early graph prompt tuning approaches relied on task-specific designs for Graph Neural Networks (GNNs), limiting their adaptability across diverse pre-training strategies. In contrast, another promising line of research has investigated universal graph prompt tuning, which operates directly in the input graph's feature space and builds a theoretical foundation that universal graph prompt tuning can theoretically achieve an equivalent effect of any prompting function, eliminating dependence on specific pre-training strategies. Recent works propose selective node-based graph prompt tuning to pursue more ideal prompts. However, we argue that selective node-based graph prompt tuning inevitably compromises the theoretical foundation of universal graph prompt tuning. In this paper, we strengthen the theoretical foundation of universal graph prompt tuning by introducing stricter constraints, demonstrating that adding prompts to all nodes is a necessary condition for achieving the universality of graph prompts. To this end, we propose a novel model and paradigm, Learning and Editing Universal GrAph Prompt Tuning (LEAP), which preserves the theoretical foundation of universal graph prompt tuning while pursuing more ideal prompts. Specifically, we first build the basic universal graph prompts to preserve the theoretical foundation and then employ actor-critic reinforcement learning to select nodes and edit prompts. Extensive experiments on graph- and node-level tasks across various pre-training strategies in both full-shot and few-shot scenarios show that LEAP consistently outperforms fine-tuning and other prompt-based approaches.

Figures (10)

• Figure 1: Different paradigms for universal graph prompt tuning and basic fine-tuning. a) Fine-tuning, b) Conventional Graph Prompt Tuning, c) Selective Node-based Prompt Tuning, and d) Learning and Editing Graph Prompt Tuning.
• Figure 2: The ROC-AUC curves for the training and testing sets.
• Figure 3: The average ROC-AUC across all datasets for the graph-level task under different hyper-parameter $k$, with the circle marking the optimal results.
• Figure 4: The average Accuracy and Macro F1 across all datasets for the node-level task under different hyper-parameter $k$, with the circle marking the optimal results.
• Figure 5: The average ROC-AUC across all datasets for the graph-level task under different hyper-parameter $\vartheta$, with the circle marking the optimal results.

Theorems & Definitions (3)

• Remark 1
• Theorem 1
• Remark 2