A new survey paper published on arXiv details the advancements in Graph Transformers (GTs), a type of model that enhances graph neural networks by addressing limitations like over-smoothing and over-squashing. The paper categorizes GT architectures based on their methods for processing structural information, including graph tokenization and structure-aware attention. It also examines the theoretical expressivity of these models and contrasts them with other graph learning algorithms. The survey organizes applications by relational, geometric, dynamic, and heterogeneous graph forms, offering practical guidance for practitioners on selecting appropriate architectural components for given input structures, while also discussing current challenges and future research directions. AI
IMPACT Provides a structured overview of Graph Transformer architectures and applications, aiding researchers and practitioners in understanding and developing graph-based AI models.
RANK_REASON This is a survey paper detailing existing research on Graph Transformers, not a novel model release or significant industry event. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX Code Finder for Papers
- Chaohao Yuan
- Connected Papers
- CORE Recommender
- DagsHub
- Gotit.pub
- graph neural networks
- Graph Transformers
- Hugging Face
- IArxiv Recommender
- Litmaps
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
