GAUGE is a pretrainable Riemannian graph model with neural vector bundles and a Dirichlet loss that captures transferable intrinsic geometry, validated on zero-shot link prediction and graph isomorphism.
hub
Deep graph contrastive representation learning.arXiv preprint arXiv:2006.04131
24 Pith papers cite this work. Polarity classification is still indexing.
24
Pith papers citing it
hub tools
representative citing papers
Feature reconstruction in GSSL is robust to noise in text-driven biomedical graphs while relation reconstruction is sensitive, with bidirectional GNN architectures performing better on noisy data and yielding up to 7% gains over language model baselines.
NodePFN pre-trains on synthetic graphs with controllable homophily and causal feature-label models to achieve 71.27 average accuracy on 23 node classification benchmarks without graph-specific training.
HyperGRL places graph nodes on a hypersphere and minimizes Helmholtz free energy with structural binding energy and mean-field repulsive potential, regulated by an adaptive thermostat, to produce discriminative representations.
UWE is a task-agnostic bi-encoder that uses many-to-many InfoNCE and token-level soft late interaction to achieve zero-shot ranking across unseen work-related target spaces while using far fewer parameters than Qwen3-8B and improving MAP by 4.4 points.
A new multi-view graph contrastive learning method generates adaptive views via learnable fractional-order diffusion dynamics instead of manual augmentations and outperforms prior baselines.
CP-GBA distills a queryable repository of promptable subgraph triggers via graph prompt learning to achieve transferable backdoor attacks on GNNs with state-of-the-art success rates across paradigms and defenses.
SAOT applies structure-aware optimal transport to capture global inter-node correspondences and uses cross-task distillation to retain prior structural knowledge, yielding accuracy gains of up to 15% on Products-CL in class-incremental settings.
HyRAG improves zero-shot generalization of graph foundation models by indexing and retrieving from tree-structured knowledge in hyperbolic space via multi-granularity retrieval and dual-path fusion.
L2IR uses LLMs to extract latent intents from behaviors and connections, improving graph fraud detection under camouflage via adaptive self-training and serving as a plug-in for GNN detectors with up to 8.27% AUPRC gain.
S2Aligner decouples semantic and structural components in LLM-as-Aligner pre-training for sparse TAGs and uses structure-oriented reconstruction plus domain risk balancing to improve transferability and reduce generalization gaps.
GCD-FGL mitigates neighborhood absorption and global semantic inconsistency in federated generalized category discovery, delivering +4.86 average HRScore gain over baselines on five graph datasets.
SDM-SCR uses LLMs for semantic disentanglement of signal from noise in text-attributed graphs followed by spectral consistency regularization to improve contrastive learning performance.
Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.
Message passing trivializes positive sample maximization in GCL via Dirichlet energy smoothing; SPGCL mitigates this by propagating only high-energy features and using low-energy ones for positive sampling.
SPG is a graph foundation model using spectral decomposition via Chebyshev filters and Gromov-Wasserstein prototypes for improved cross-graph transferability.
Contrastive FUSE learns node embeddings from partial pairwise supervision and structural signals alone by optimizing a spectral contrastive objective with a lightweight modularity approximation, yielding competitive performance and runtime gains on citation and co-purchase graphs.
CHoE uses structure-conditioned experts, structure-aware routing with load balancing, and prompt-based semantic fusion to improve few-shot performance on cross-domain heterogeneous graph prompt learning tasks.
GP2F is a dual-branch graph prompting framework that fuses frozen pre-trained knowledge with task-specific adaptation to reduce estimation error and outperform baselines in cross-domain few-shot node and graph classification.
OpenGLT benchmark finds no single GNN architecture dominates graph-level tasks, with subgraph-based models strongest in expressiveness, graph learning and SSL models in robustness, node and pooling models in efficiency, and graph topology partially guiding architecture choice.
DiGGR introduces a self-supervised graph representation learning framework that disentangles latent factors to guide mask modeling and improve representation quality on graph tasks.
Empirical comparison shows gradient-based explanations for GNN node similarities are actionable, consistent, and retain effects when sparsified, unlike mutual information explanations.
MPAIACL applies contrastive learning to generate adversarial invariant augmentations that improve GNN generalization under covariate shifts on graphs.
BrainSimSiam applies positive-only Siamese self-supervised learning to fMRI data to produce representations that generalize across downstream tasks and outperform supervised baselines.
citing papers explorer
-
Are Common Substructures Transferable? Riemannian Graph Foundation Model with Neural Vector Bundles
GAUGE is a pretrainable Riemannian graph model with neural vector bundles and a Dirichlet loss that captures transferable intrinsic geometry, validated on zero-shot link prediction and graph isomorphism.
-
Robustness of Graph Self-Supervised Learning to Real-World Noise: A Case Study on Text-Driven Biomedical Graphs
Feature reconstruction in GSSL is robust to noise in text-driven biomedical graphs while relation reconstruction is sensitive, with bidirectional GNN architectures performing better on noisy data and yielding up to 7% gains over language model baselines.
-
Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors
NodePFN pre-trains on synthetic graphs with controllable homophily and causal feature-label models to achieve 71.27 average accuracy on 23 node classification benchmarks without graph-specific training.
-
Energy-Balanced Hyperspherical Graph Representation Learning via Structural Binding and Entropic Dispersion
HyperGRL places graph nodes on a hypersphere and minimizes Helmholtz free energy with structural binding energy and mean-field repulsive potential, regulated by an adaptive thermostat, to produce discriminative representations.
-
Unified Work Embeddings: Contrastive Learning of a Bidirectional Multi-task Ranker
UWE is a task-agnostic bi-encoder that uses many-to-many InfoNCE and token-level soft late interaction to achieve zero-shot ranking across unseen work-related target spaces while using far fewer parameters than Qwen3-8B and improving MAP by 4.4 points.
-
Adaptive Multi-view Graph Contrastive Learning via Fractional-order Neural Diffusion Networks
A new multi-view graph contrastive learning method generates adaptive views via learnable fractional-order diffusion dynamics instead of manual augmentations and outperforms prior baselines.
-
Cross-Paradigm Graph Backdoor Attacks with Promptable Subgraph Triggers
CP-GBA distills a queryable repository of promptable subgraph triggers via graph prompt learning to achieve transferable backdoor attacks on GNNs with state-of-the-art success rates across paradigms and defenses.
-
SAOT: Self-Supervised Continual Graph Learning with Structure-Aware Optimal Transport
SAOT applies structure-aware optimal transport to capture global inter-node correspondences and uses cross-task distillation to retain prior structural knowledge, yielding accuracy gains of up to 15% on Products-CL in class-incremental settings.
-
Generalizing Graph Foundation Models via Hyperbolic Retrieval-Augmented Generation
HyRAG improves zero-shot generalization of graph foundation models by indexing and retrieving from tree-structured knowledge in hyperbolic space via multi-granularity retrieval and dual-path fusion.
-
L2IR: Revealing Latent Intent in Graph Fraud Detection
L2IR uses LLMs to extract latent intents from behaviors and connections, improving graph fraud detection under camouflage via adaptive self-training and serving as a plug-in for GNN detectors with up to 8.27% AUPRC gain.
-
S2Aligner: Pair-Efficient and Transferable Pre-Training for Sparse Text-Attributed Graphs
S2Aligner decouples semantic and structural components in LLM-as-Aligner pre-training for sparse TAGs and uses structure-oriented reconstruction plus domain risk balancing to improve transferability and reduce generalization gaps.
-
Generalized Category Discovery in Federated Graph Learning
GCD-FGL mitigates neighborhood absorption and global semantic inconsistency in federated generalized category discovery, delivering +4.86 average HRScore gain over baselines on five graph datasets.
-
Disentangle-then-Refine: LLM-Guided Decoupling and Structure-Aware Refinement for Graph Contrastive Learning
SDM-SCR uses LLMs for semantic disentanglement of signal from noise in text-attributed graphs followed by spectral consistency regularization to improve contrastive learning performance.
-
Revisiting Positive Samples in Graph Contrastive Learning: From the Perspective of Message Passing
Message passing trivializes positive sample maximization in GCL via Dirichlet energy smoothing; SPGCL mitigates this by propagating only high-energy features and using low-energy ones for positive sampling.
-
A Graph Foundation Model with Spectral Parsing and Prototype-Guided Spatial Propagation
SPG is a graph foundation model using spectral decomposition via Chebyshev filters and Gromov-Wasserstein prototypes for improved cross-graph transferability.
-
Fast and Featureless Node Representation Learning with Partial Pairwise Supervision
Contrastive FUSE learns node embeddings from partial pairwise supervision and structural signals alone by optimizing a spectral contrastive objective with a lightweight modularity approximation, yielding competitive performance and runtime gains on citation and co-purchase graphs.
-
CHoE: Cross-Domain Heterogeneous Graph Prompt Learning via Structure-Conditioned Experts
CHoE uses structure-conditioned experts, structure-aware routing with load balancing, and prompt-based semantic fusion to improve few-shot performance on cross-domain heterogeneous graph prompt learning tasks.
-
GP2F: Cross-Domain Graph Prompting with Adaptive Fusion of Pre-trained Graph Neural Networks
GP2F is a dual-branch graph prompting framework that fuses frozen pre-trained knowledge with task-specific adaptation to reduce estimation error and outperform baselines in cross-domain few-shot node and graph classification.
-
OpenGLT: A Comprehensive Benchmark of Graph Neural Networks for Graph-Level Tasks
OpenGLT benchmark finds no single GNN architecture dominates graph-level tasks, with subgraph-based models strongest in expressiveness, graph learning and SSL models in robustness, node and pooling models in efficiency, and graph topology partially guiding architecture choice.
-
Disentangled Generative Graph Representation Learning
DiGGR introduces a self-supervised graph representation learning framework that disentangles latent factors to guide mask modeling and improve representation quality on graph tasks.
-
Explaining Graph Neural Networks for Node Similarity on Graphs
Empirical comparison shows gradient-based explanations for GNN node similarities are actionable, consistent, and retain effects when sparsified, unlike mutual information explanations.
-
Graph Data Augmentation with Contrastive Learning on Covariate Distribution Shift
MPAIACL applies contrastive learning to generate adversarial invariant augmentations that improve GNN generalization under covariate shifts on graphs.
-
Learning Robust and Task-Invariant Functional Representation from fMRI through Siamese Self-Supervised Learning
BrainSimSiam applies positive-only Siamese self-supervised learning to fMRI data to produce representations that generalize across downstream tasks and outperform supervised baselines.