The paper shows that a plain Transformer can serve as an effective graph learner by adding three lightweight modifications: simplified L₂ attention, adaptive RMS normalization, and an MLP-based positional encoding stem. These changes preserve token magnitude and enable the model to achieve high expressivity on graph benchmarks, outperforming more complex graph transformer variants. The results suggest that plain Transformers can act as a unified backbone for multimodal learning across language, vision, and graph domains.
By Liheng Ma, Soumyasundar Pal, Yingxue Zhang, Philip H. S. Torr, Mark Coates
Positional encodings (PEs) are essential for Transformers. Yet designing effective PEs for non-Euclidean graphs remains challenging.
arXiv:2602.15239v3 Announce Type: replace
Abstract: Transformers have achieved remarkable success across domains, motivating the rise of Graph Transformers (GTs) as attention-based architectures for...
By Javier Porras-Valenzuela, Zhiyang Wang, Teresa Shang, Yusu Wang, Alejandro Ribeiro
arXiv:2606. 25293v1 Announce Type: new Abstract: Positional encodings (PEs) are essential for Transformers.
By Yipeng Zhang, Zhongtian Sun, Pietro Li\`o, Kelin Xia
HyPE-GT introduces a framework that generates learnable hyperbolic positional encodings for Graph Transformers, enabling the capture of complex hierarchical relationships in graph-structured data. Unlike traditional Euclidean encodings, HyPE’s hyperbolic encodings can be selected to suit specific downstream tasks and help mitigate oversmoothing in deep Graph Neural Networks. Experiments on molecular benchmarks and large-scale Open Graph Benchmark datasets demonstrate improved performance, while additional tests on Coauthor and Copurchase networks confirm HyPE’s effectiveness in controlling oversmoothing.
By Kushal Bose, Swagatam Das
The paper investigates how the choice of graph tokenization affects transformer expressivity. It analyzes three tokenization families—spectral, random‑walk, and adjacency—showing that each induces different depth requirements and that some tokenizations are inherently lossy or ill‑conditioned for certain tasks. The authors prove lower bounds and impossibility results for converting between tokenizations and validate these findings with experiments on synthetic and real‑world data.
By Maya Bechler-Speicher, Gilad Yehudai, Gil Harari, Clayton Sanford, Amir Globerson, Joan Bruna