mlsanityData to 14 Oct 2026

GraphShaper: Geometry-aware Alignment for Improving Transfer Learning in Text-Attributed Graphs

arXiv 2510.12085 ↗First posted 14 Oct 2025cs.LG cs.GRcited by 0

Not on the map. Why?

Abstract (arXiv)

Graph foundation models represent a transformative paradigm for learning transferable representations across diverse graph domains. Recent methods leverage large language models to unify graph and text modalities into a shared representation space using contrastive learning. However, systematic evaluations reveal significant performance degradation at structural boundaries where distinct topological patterns converge, with accuracy losses exceeding 20 percentage points. This issue arises from a key limitation: current methods assume all graph structures can be encoded within a single Euclidean space. In reality, tree structures require hyperbolic geometry to preserve hierarchical branching, while cyclic patterns depend on spherical geometry for closure properties. At structural boundaries, nodes experience conflicting geometric constraints that uniform encoding spaces cannot resolve. This raises a crucial challenge: Can alignment frameworks be designed to respect the intrinsic geometric diversity of graph structures? We introduce GraphShaper, a geometry-aware framework that enhances graph encoding through multi-geometric specialization. Our approach employs expert networks tailored to different geometric spaces, dynamically computing fusion weights to adaptively integrate geometric properties based on local structural characteristics. This adaptive fusion preserves structural integrity before alignment with text embeddings. Extensive experiments demonstrate that GraphShaper achieves 9.47% accuracy improvements on citation networks and 7.63% on social networks in zero-shot settings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 35 of its 51 references that are in the map. First posted on arXiv 14 Oct 2025.

Builds on

35 of 51 references.

References · 35

  1. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  2. 2018Deep Graph Infomaxcited by 1,106
  3. 2016Revisiting Semi-Supervised Learning with Graph Embeddingscited by 976
  4. 2015Image-based Recommendations on Styles and Substitutescited by 706
  5. 2020Contrastive Multi-View Representation Learning on Graphscited by 652
  6. 2020Graph Contrastive Learning with Adaptive Augmentationcited by 519
  7. 2020GCC: Graph Contrastive Coding for Graph Neural Network Pre-Trainingcited by 463
  8. 2021Graph Contrastive Learning Automatedcited by 287
  9. 2023Exploring the Potential of Large Language Models (LLMs) in Learning on Graphscited by 274
  10. 2023Can Language Models Solve Graph Problems in Natural Language?cited by 252
  11. 2023One for All: Towards Training One Graph Model for All Classification Taskscited by 243
  12. 2023Language is All a Graph Needscited by 220
23 more
  1. 2023GPT4Graph: Can Large Language Models Understand Graph Structured Data ? An Empirical Evaluation and Benchmarkingcited by 210
  2. 2023NodeFormer: A Scalable Graph Structure Learning Transformer for Node Classificationcited by 210
  3. 2024LLaGA: Large Language and Graph Assistantcited by 181
  4. 2022Learning on Large-scale Text-attributed Graphs via Variational Inferencecited by 155
  5. 2023Representation Learning with Large Language Models for Recommendationcited by 155
  6. 2024Let Your Graph Do the Talking: Encoding Structured Data for LLMscited by 118
  7. 2023GraphText: Graph Reasoning in Text Spacecited by 111
  8. 2024ZeroG: Investigating Cross-dataset Zero-shot Transferability in Graphscited by 85
  9. 2023Augmenting Low-Resource Text Classification with Graph-Grounded Pre-training and Promptingcited by 62
  10. 2024LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddingscited by 61
  11. 2024GraphInstruct: Empowering Large Language Models with Graph Understanding and Reasoning Capabilitycited by 58
  12. 2024Efficient Tuning and Inference for Large Language Models on Textual Graphscited by 54
  13. 2024GLBench: A Comprehensive Benchmark for Graph with Large Language Modelscited by 36
  14. 2025RiemannGFM: Learning a Graph Foundation Model from Riemannian Geometrycited by 35
  15. 2023ConGraT: Self-Supervised Contrastive Pretraining for Joint Graph and Text Embeddingscited by 32
  16. 2024GraphWiz: An Instruction-Following Language Model for Graph Problemscited by 32
  17. 2025LLM as GNN: Graph Vocabulary Learning for Text-Attributed Graph Foundation Modelscited by 22
  18. 2024GraphInsight: Unlocking Insights in Large Language Models for Graph Structure Understandingcited by 18
  19. 2025Collaboration of Large Language Models and Small Recommendation Models for Device-Cloud Recommendationcited by 15
  20. 2024Graph Learning in the Era of LLMs: A Survey from the Perspective of Data, Models, and Taskscited by 10
  21. 2024A Survey on Self-Supervised Graph Foundation Models: Knowledge-Based Perspectivecited by 9
  22. 2025Toward General and Robust LLM-enhanced Text-attributed Graph Learningcited by 6
  23. 2024UniGAP: A Universal and Adaptive Graph Upsampling Approach to Mitigate Over-Smoothing in Node Classification Taskscited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Surprisingly Simple and Effective Multi-Domain Graph Foundation Model through Graph-to-Table Alignment
  2. 2025SA2GFM: Enhancing Robust Graph Foundation Models with Structure-Aware Semantic Augmentation
  3. 2025Align-GRAG: Anchor and Rationale Guided Dual Alignment for Graph Retrieval-Augmented Generation
  4. 2025Quantizing Text-attributed Graphs for Semantic-Structural Integration
  5. 2026A Unified Graph Language Model for Multi-Domain Multi-Task Graph Alignment Instruction Tuning
  6. 2025MLaGA: Multimodal Large Language and Graph Assistant
  7. 2026Distilling Graph Geometry: Knowledge Gap from GNNs to MLPs
  8. 2024GUNDAM: Aligning Large Language Models with Graph Understanding