Graph Foundation Models
PulseAugur coverage of Graph Foundation Models — every cluster mentioning Graph Foundation Models across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New STAG framework enables stealthy backdoor attacks on graph foundation models
Researchers have developed STAG, a novel framework designed to stealthily inject backdoors into Graph Foundation Models (GFMs) that process text-attributed graphs (TAGs). Unlike previous attacks that target graph or tex…
-
New MSB-GFM framework tackles multi-label node classification across domains
Researchers have introduced the Multi-Semantic Basis Graph Foundation Model (MSB-GFM), a novel framework designed to improve cross-domain multi-label node classification. Current Graph Foundation Models (GFMs) struggle …
-
New ProGFM model enhances knowledge transfer for graph foundation models
Researchers have introduced ProGFM, a novel Propagation-aware Graph Foundation Model designed to enhance knowledge transfer across diverse graph domains. Unlike previous models that focused on feature and structure alig…
-
New SliGFM Model Unifies Heterogeneous Node Features for Graph Foundation Models
Researchers have introduced SliGFM, a novel graph foundation model designed to address the challenge of unifying heterogeneous node features across diverse graph domains. The model is built on principles of formal unifo…
-
New LLM methods enable dynamic graph reasoning with agent-like nodes
Researchers have introduced two novel approaches to enhance large language models' (LLMs) ability to reason with graph data. The first, 'agentic graph token reasoning,' allows LLMs to dynamically select and encode graph…
-
New multimodal graph foundation model CHARM enables zero-shot transfer learning
Researchers have introduced CHARM, a novel multimodal graph foundation model designed for zero-shot transfer learning on complex graph datasets. CHARM addresses the challenges of generalizing knowledge across different …
-
New attack vector targets graph foundation models via alignment layer
Researchers have identified a new attack vector targeting graph foundation models by exploiting their alignment layer, which maps inputs into a shared representation. This attack, performed at inference time without acc…
-
Graph Transformers for MILPs Limited by 1-WL Test, Study Finds
A new paper characterizes the expressive power of global-attention graph transformers used for mixed-integer linear programs (MILPs). The research proves that these models, including architectures like Graphormer and Se…
-
New GTAlign Framework Simplifies Graph Foundation Models
Researchers have introduced GTAlign, a novel framework for creating text-free Graph Foundation Models (GFMs). This approach aims to bridge the gap between graph topology and tabular representation spaces, enabling GFMs …
-
New study finds advanced GFMs only slightly outperform GNNs on node prediction tasks
A recent study re-evaluated nine Graph Foundation Models (GFMs) for node property prediction tasks, a common application in Graph ML used for areas like fraud detection and recommendation systems. The research found tha…
-
New framework enables Graph Foundation Models for network dynamics
Researchers have introduced a new framework for Graph Foundation Models (GFMs) designed to handle network dynamics across different systems. Their approach, demonstrated by a model called ts-net, shows zero-shot general…
-
New framework reveals geometry-dependent performance in relational learning models
Researchers have introduced a new framework for evaluating relational learning models, moving beyond standard leaderboards that average performance across diverse datasets. This new approach stratifies datasets by their…
-
New theory quantifies graph model adaptation, introduces Message Tuning
Researchers have introduced Prismatic Space Theory (PS-Theory) to quantify the adaptation capacity of methods used for Graph Foundation Models (GFMs). This framework establishes an upper bound for graph prompt tuning, a…
-
New HyRAG framework boosts graph model generalization
Researchers have developed a new framework called Hyperbolic Retrieval-Augmented Generation (HyRAG) to improve the generalization capabilities of Graph Foundation Models (GFMs). Existing RAG methods struggle with the ge…
-
GFMate enhances Graph Foundation Models with test-time prompt tuning
Researchers have introduced GFMate, a novel test-time prompt tuning method designed to enhance Graph Foundation Models (GFMs). Unlike previous approaches that embed source-domain information into prompts, GFMate applies…
-
New research advances graph representation learning and Shapley value computation
Researchers are developing advanced methods for graph representation learning, focusing on improving generalization and efficiency. New models like SPG aim to parse spectral responses and use prototype-guided propagatio…
-
New theories explore graph foundation model transferability and cross-modal learning
Two new research papers explore the challenges and potential of graph foundation models (GFMs). The first paper, "When Do Graph Foundation Models Transfer? A Data-Centric Theory," investigates the properties of graph do…
-
New Graph Foundation Model Learns Multi-Scale Representations
Researchers have introduced R-GFM, a novel Graph Foundation Model that utilizes a Riemannian Graph-of-Graphs approach to address limitations in existing models. Unlike previous methods that use fixed-hop subgraph sampli…
-
New diagnostics for graph models assess structural learning vs. node features
Researchers have introduced a new diagnostic framework for graph foundation models using graph invariants. This approach aims to disentangle the impact of node features from graph structure in benchmark evaluations. The…