SGL-PT: A Strong Graph Learner with Graph Prompt Tuning

Recently, much exertion has been paid to design graph self-supervised methods to obtain generalized pre-trained models, and adapt pre-trained models onto downstream tasks through fine-tuning. However, there exists an inherent gap between pretext and downstream graph tasks, which insufficiently exerts the ability of pre-trained models and even leads to negative transfer. Meanwhile, prompt tuning has seen emerging success in natural language processing by aligning pre-training and fine-tuning with consistent training objectives. In this paper, we identify the challenges for graph prompt tuning: The first is the lack of a strong and universal pre-training task across sundry pre-training methods in graph domain. The second challenge lies in the difficulty of designing a consistent training objective for both pre-training and downstream tasks. To overcome above obstacles, we propose a novel framework named SGL-PT which follows the learning strategy ``Pre-train, Prompt, and Predict''. Specifically, we raise a strong and universal pre-training task coined as SGL that acquires the complementary merits of generative and contrastive self-supervised graph learning. And aiming for graph classification task, we unify pre-training and fine-tuning by designing a novel verbalizer-free prompting function, which reformulates the downstream task in a similar format as pretext task. Empirical results show that our method surpasses other baselines under unsupervised setting, and our prompt tuning method can greatly facilitate models on biological datasets over fine-tuning methods.

Summary Notes

Revolutionizing Graph Neural Networks: An Introduction to SGL-PT

Identifying the Challenges

• Need for Universal Pre-training Tasks: Traditional graph self-supervised learning methods focus more on intra-data (local) relations, leaving inter-data (global) relations between graphs less explored. SGL-PT addresses this by integrating generative and contrastive models to capture a broader spectrum of graph relations.

• Reformulating Downstream Tasks: Adapting pre-trained models to downstream tasks is often challenging due to format discrepancies. SGL-PT simplifies this process, ensuring a smoother transition between pre-training and application phases.

Core Components of SGL-PT

Local and Global Branch Integration

• Local Branch: Focuses on enhancing intra-data relations through a graph masked autoencoder approach, improving relational reasoning and robustness.

• Global Branch: Aims at better understanding inter-data relations by employing graph contrastive learning, furthering the model's discriminatory capabilities.

• Integration: Combines the strengths of both branches, leveraging robustness and discriminative features for a comprehensive learning approach.

Advancements in Prompt-based Learning

• Task Reformulation: Introduces a masked super node that simplifies adapting the pre-trained model to downstream tasks, effectively summarizing the graph's information.

• Class Mapping Without Verbalizers: Utilizes supervised graph prototypical contrastive learning to streamline class mapping, eliminating the need for explicit verbalizer design.

Empirical Success

Building on Previous Works

Conclusion: The Future of GNNs with SGL-PT

How Athina AI can help