H4G: Unlocking Faithful Inference for Zero-Shot Graph Learning in Hyperbolic Space
Not on the map. Why?
Abstract (arXiv)
Text-attributed graphs are widely used across domains, offering rich opportunities for zero-shot learning via graph-text alignment. However, existing methods struggle with tasks requiring fine-grained pattern recognition, particularly on heterophilic graphs. Through empirical and theoretical analysis, we identify an over-abstraction problem: current approaches operate at excessively large hyperbolic radii, compressing multi-scale structural information into uniform high-level abstractions. This abstraction-induced information loss obscures critical local patterns essential for accurate predictions. By analyzing embeddings in hyperbolic space, we demonstrate that optimal graph learning requires faithful preservation of fine-grained structural details, better retained by representations positioned closer to the origin. To address this, we propose H4G, a framework that systematically reduces embedding radii using learnable block-diagonal scaling matrices and M"obius matrix multiplication. This approach restores access to fine-grained patterns while maintaining global receptive ability with minimal computational overhead. Experiments show H4G achieves state-of-the-art zero-shot performance with 12.8% improvement on heterophilic graphs and 8.4% on homophilic graphs, confirming that radius reduction enables faithful multi-scale representation for advancing zero-shot graph learning.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 22 of its 52 references that are in the map. First posted on arXiv 14 Oct 2025.
Builds on
22 of 52 references.
References · 22
- 2020Open Graph Benchmark: Datasets for Machine Learning on Graphscited by 2,148
- 2016Revisiting Semi-Supervised Learning with Graph Embeddingscited by 970
- 2015Image-based Recommendations on Styles and Substitutescited by 702
- 2023One for All: Towards Training One Graph Model for All Classification Taskscited by 239
- 2023Language is All a Graph Needscited by 219
- 2024LLaGA: Large Language and Graph Assistantcited by 177
- 2023GraphPrompt: Unifying Pre-Training and Downstream Tasks for Graph Neural Networkscited by 166
- 2023GraphText: Graph Reasoning in Text Spacecited by 110
- 2020Graph Prototypical Networks for Few-shot Learning on Attributed Networkscited by 90
- 2024ZeroG: Investigating Cross-dataset Zero-shot Transferability in Graphscited by 83
- 2024Fine-tuning large language models for domain adaptation: Exploration of training strategies, scaling, model merging and synergistic capabilitiescited by 67
- 2024LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddingscited by 60
10 more
- 2024GraphCLIP: Enhancing Transferability in Graph Foundation Models for Text-Attributed Graphscited by 58
- 2024GraphInstruct: Empowering Large Language Models with Graph Understanding and Reasoning Capabilitycited by 57
- 2024Efficient Tuning and Inference for Large Language Models on Textual Graphscited by 54
- 2023ConGraT: Self-Supervised Contrastive Pretraining for Joint Graph and Text Embeddingscited by 32
- 2024GraphInsight: Unlocking Insights in Large Language Models for Graph Structure Understandingcited by 18
- 2024A Survey of Few-Shot Learning on Graphs: from Meta-Learning to Pre-Training and Prompt Learningcited by 5
- 2024Tree-Wasserstein Distance for High Dimensional Data with a Latent Feature Hierarchycited by 5
- 2023Graph Contrastive Learning Meets Graph Meta Learning: A Unified Method for Few-shot Node Taskscited by 3
- 2025Curvature Learning for Generalization of Hyperbolic Neural Networkscited by 3
- 2024A Geometry-Aware Algorithm to Learn Hierarchical Embeddings in Hyperbolic Spacecited by 2
Built on by
Citing papers per year: 2026: 1.
1 paper.