Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddings

About

Zero-shot graph machine learning, especially with graph neural networks (GNNs), has garnered significant interest due to the challenge of scarce labeled data. While methods like self-supervised learning and graph prompt learning have been extensively explored, they often rely on fine-tuning with task-specific labels, limiting their effectiveness in zero-shot scenarios. Inspired by the zero-shot capabilities of instruction-fine-tuned large language models (LLMs), we introduce a novel framework named Token Embedding-Aligned Graph Language Model (TEA-GLM) that leverages LLMs as cross-dataset and cross-task zero-shot learners for graph machine learning. Concretely, we pretrain a GNN, aligning its representations with token embeddings of an LLM. We then train a linear projector that transforms the GNN's representations into a fixed number of graph token embeddings without tuning the LLM. A unified instruction is designed for various graph tasks at different levels, such as node classification (node-level) and link prediction (edge-level). These design choices collectively enhance our method's effectiveness in zero-shot learning, setting it apart from existing methods. Experiments show that our graph token embeddings help the LLM predictor achieve state-of-the-art performance on unseen datasets and tasks compared to other methods using LLMs as predictors.

Duo Wang, Yuan Zuo, Fengzhi Li, Junjie Wu• 2024

Related benchmarks

TaskDatasetResultRank
Graph ClassificationPROTEINS
Accuracy67.88
1383
Graph ClassificationMUTAG
Accuracy71.1
1229
Node ClassificationCora (test)
Mean Accuracy62.38
951
Node ClassificationPubmed--
902
Node ClassificationTexas
Accuracy0.791
859
Node ClassificationPubMed (test)
Accuracy84.8
628
Node ClassificationPubmed
Accuracy86.6
501
Node ClassificationPhoto
Mean Accuracy71.6
447
Node ClassificationarXiv
Accuracy65.5
325
Node ClassificationREDDIT
Accuracy54.7
268
Showing 10 of 95 rows
...

Other info

Follow for update