graph neural networks
PulseAugur coverage of graph neural networks — every cluster mentioning graph neural networks across labs, papers, and developer communities, ranked by signal.
- instance of Graph Neural Networks (GNNs) 95%
- developed alphaXiv 90%
- instance of alphaXiv 90%
- developed Gotit.pub 90%
- instance of CatalyzeX 90%
- instance of IArxiv 90%
- instance of graph attention network 90%
- instance of Graphsage 90%
- instance of MPNNs 90%
- instance of Graph Convolutional Networks 90%
- used by Link prediction 80%
- used by Node Classification 80%
- 2026-05-25 research_milestone Researchers proposed new polynomial-time algorithms for explaining Graph Neural Networks. source
- 2026-05-13 research_milestone A new graph neural network architecture was introduced for the multicut problem. source
- 2026-05-11 research_milestone A new method for pre-training GNNs using ECFPs shows improved performance in QSAR tasks. source
21 day(s) with sentiment data
-
Neumathway project merges GNNs with symbolic engines for hallucination-free AI math
An open-source project called Neumathway is being developed to combine graph neural networks (GNNs) with symbolic engines. This approach aims to create AI systems that can perform mathematical reasoning without generati…
-
New RL approach enhances graph optimization with projection agents
Researchers have developed a new methodology called "projection agents" to improve the performance of reinforcement learning (RL) models in graph combinatorial optimization (GCO) problems. This approach, built on the Wo…
-
arXiv survey details traditional and deep learning methods for infrared small target detection · 2 sources tracked
A two-part survey paper published on arXiv details advancements in infrared small target detection. Part 1 covers traditional methods, including model-based approaches like low-rank and sparse decomposition, alongside m…
-
LLM agent RelAgent acts as autonomous data scientist for relational learning
Researchers have introduced RelAgent, an LLM-based autonomous data scientist designed for relational learning tasks. This system operates in two phases: a search phase where an LLM agent constructs SQL feature programs …
-
New method tackles under-confidence in Graph Neural Networks
Researchers have developed a new method to address under-confidence in Graph Neural Networks (GNNs). This approach establishes a theoretical framework that links model confidence to both class-centroid and node-level ca…
-
Graph Neural Networks poised to advance astrophysics and cosmology research
A new paper explores the application of Graph Neural Networks (GNNs) in astrophysics and cosmology, arguing that their ability to model relationships between objects makes them well-suited for physical data. The review …
-
New research enhances LLM graph reasoning and retrieval; software updates released
Recent research explores advanced graph reasoning and retrieval techniques for large language models. One paper introduces Federated GraphRAG (FedGRAG) to enable distributed knowledge graphs for LLMs without compromisin…
-
New GNN framework SOCPNet learns iterative solutions for optimization problems
Researchers have developed SOCPNet, a novel graph neural network (GNN) framework designed to solve second-order cone programs (SOCPs). Unlike previous GNN approaches that predict solutions in a single pass, SOCPNet lear…
-
LLMSeg pipeline uses LLMs for image segmentation without training
A new training-free pipeline called LLMSeg has been developed that leverages multimodal large language models (MLLMs) for image segmentation. This approach bypasses the need for dedicated segmentation architectures or t…
-
New method RAW-Explainer enhances interpretability of GNNs for knowledge graphs
Researchers have developed RAW-Explainer, a novel method for interpreting the predictions of path reasoning graph neural networks (GNNs) used in knowledge graph completion. This technique generates concise and connected…
-
New diagnostic tool probes GNN failure modes
Researchers have developed a new diagnostic tool called Layer-Wise Probe Entropy to analyze the internal workings of graph neural networks (GNNs). This method examines how task-relevant information evolves through each …
-
New Graph Homomorphism Distortion Metric Enhances Graph Neural Networks
Researchers have introduced a new metric called Graph Homomorphism Distortion to better understand the relationship between structure and features in graph learning. This metric aims to quantify how node features are af…
-
New CostStart method halves solve times for simplex algorithm
Researchers have developed a new method called CostStart to improve the efficiency of the simplex algorithm for solving linear programs. While previous approaches using graph neural networks reduced iteration counts, th…
-
AMIGA research disentangles graph learning features and topology
A new research paper introduces AMIGA, a novel approach to Graph Additive Models designed to disentangle feature, topology, and interaction effects in graph learning. AMIGA addresses limitations in existing Graph Neural…
-
New benchmark Lethe evaluates edge unlearning in graph neural networks
Researchers have introduced Lethe, a new benchmark designed to evaluate edge unlearning methods for graph neural networks. Current evaluations for edge unlearning are limited, often using datasets where node features ov…
-
ChronoSheaf network enhances graph neural networks with spatio-temporal filtering
Researchers have introduced ChronoSheaf, a novel spatio-temporal network designed to enhance graph neural networks by incorporating temporal dynamics. This new architecture utilizes Transitional Restriction Maps (TRM) a…
-
Graph databases don't reveal hidden data but shift computational work, study finds
A recent analysis explored the utility of graph databases by comparing them to document stores for data analysis. The study found that for known questions, graph databases do not reveal new insights beyond what is alrea…
-
CoViG enhances Vision Graph efficiency for image classification
Researchers have introduced CoViG, a new method that enhances the efficiency of Vision Graphs (ViGs) for image classification. By incorporating a MinCUT pooling step within the Grapher component of ViG blocks, CoViG sim…
-
Simple baselines match complex models in relational deep learning
Researchers have revisited simple baseline methods for relational deep learning, which represents relational databases as temporal heterogeneous graphs. Their study focused on exploiting target attributes, relational at…
-
Graph Neural Networks integrated with graph databases for faster training and inference
Researchers have explored the integration of Graph Neural Networks (GNNs) with graph databases, a novel approach to handling large graph data. The study demonstrates that graph databases can directly execute GNN trainin…