English

Graph Tokenization for Bridging Graphs and Transformers

Machine Learning 2026-03-13 v1 Artificial Intelligence

Abstract

The success of large pretrained Transformers is closely tied to tokenizers, which convert raw input into discrete symbols. Extending these models to graph-structured data remains a significant challenge. In this work, we introduce a graph tokenization framework that generates sequential representations of graphs by combining reversible graph serialization, which preserves graph information, with Byte Pair Encoding (BPE), a widely adopted tokenizer in large language models (LLMs). To better capture structural information, the graph serialization process is guided by global statistics of graph substructures, ensuring that frequently occurring substructures appear more often in the sequence and can be merged by BPE into meaningful tokens. Empirical results demonstrate that the proposed tokenizer enables Transformers such as BERT to be directly applied to graph benchmarks without architectural modifications. The proposed approach achieves state-of-the-art results on 14 benchmark datasets and frequently outperforms both graph neural networks and specialized graph transformers. This work bridges the gap between graph-structured data and the ecosystem of sequence models. Our code is available at \href{https://github.com/BUPT-GAMMA/Graph-Tokenization-for-Bridging-Graphs-and-Transformers}{\color{blue}here}.

Keywords

Cite

@article{arxiv.2603.11099,
  title  = {Graph Tokenization for Bridging Graphs and Transformers},
  author = {Zeyuan Guo and Enmao Diao and Cheng Yang and Chuan Shi},
  journal= {arXiv preprint arXiv:2603.11099},
  year   = {2026}
}

Comments

Accepted as a poster at ICLR 2026. Code is available at https://github.com/BUPT-GAMMA/Graph-Tokenization-for-Bridging-Graphs-and-Transformers

R2 v1 2026-07-01T11:15:14.416Z