GraphLoRA proposes a novel framework for Large Language Model Recommendation (LLMRec) that integrates structural information with textual semantics. It achieves this by embedding a trainable graph message-passing network within the low-rank adaptation pathway, allowing collaborative topology to explicitly guide parameter updates.
This paper proposes a multi-level temporal graph network with local-global feature fusion for industrial fault diagnosis. It addresses the complex multi-level relations among sensors by dynamically constructing correlation graphs and combining LSTM-based encoders for temporal features with graph convolution layers for spatial dependencies.
This study introduces a graph-based hierarchical reinforcement learning approach for the automated co-design of high-performance thermodynamic cycles. It encodes cycles as graphs, uses a deep learning surrogate for decoding, and employs a hierarchical RL framework for structural evolution and parameter optimization.
This research introduces `subgraph2vec`, a novel method for learning distributed representations of rooted sub-graphs extracted from large graphs. It aims to embed complex graph structures into a lower-dimensional vector space, facilitating various downstream machine learning tasks on graph data.
This paper details the use of graph neural networks (GNNs) for photovoltaic power forecasting on edge intelligent meters in a microgrid. It explores the training and deployment of GCN and GraphSAGE models, including a customized ONNX operator, with a real-world case study demonstrating successful execution on smart meters.
DBGL introduces a novel Decay-Aware Bipartite Graph Learning method to address the challenges of irregular medical time series classification. It utilizes a patient-variable bipartite graph to model irregular sampling patterns and variable relationships, alongside a node-specific temporal decay encoding for variable decay irregularity.
The ST-GAT framework provides an explainable Graph Neural Network solution for early detection of bank distress and interbank contagion surveillance in the U.S. banking sector. It models over 8,000 FDIC institutions using dynamic graphs, achieving high performance (AUPRC 0.939) and identifying key predictive factors like ROA and NPL Ratio.
Graph-MLP introduces a novel approach for node classification in graph neural networks, diverging from traditional message passing mechanisms. This method aims to improve efficiency and performance by classifying nodes directly within the graph structure.
This content discusses the use of neural embeddings to represent graphs. The research explores the application of hyperbolic space for this representation, indicating an advanced topic in machine learning.
The paper introduces Hierarchical Multi-view HAAR (HMH), a novel spectral graph-learning framework designed to address oversmoothing and degree-biased aggregation in GNNs for heterophilous graphs. HMH constructs a soft graph hierarchy and applies learnable spectral filters using sparse, orthonormal Haar bases, achieving near-linear time scalability.
This content explores the application of probabilistic graph neural inference for planetary geology survey missions. It focuses on integrating these techniques within carbon-negative infrastructure.
This content introduces the challenge of maintaining bio-inspired soft robotics, where traditional sensor methods fail due to the robots' flexible nature. It proposes a novel solution combining probabilistic graph neural networks with zero-trust architecture for effective monitoring and governance.
The author describes a eureka moment while modeling circular manufacturing supply chains using Graph Neural Networks in scenarios of extreme data sparsity. The key insight was to embrace the inherent uncertainty through probabilistic inference techniques, rather than trying to force more data into the system.
The author discusses the limitations of centralized AI for low-power coastal monitoring, where sending raw sensor data to the cloud quickly drained device batteries. This challenge inspired the exploration of Probabilistic Graph Neural Inference to enable efficient, localized reasoning for edge intelligence in autonomous deployments.
Directed Graph Convolutional Networks (DGCNs) are a specialized type of neural network designed to process data structured as directed graphs. They extend traditional GCNs by effectively capturing directional information and relationships within complex network structures.
arXiv paper 2605.11145 introduces DPAA, a debiasing framework for GNN-based collaborative filtering that applies adaptive weighting during message passing. It effectively counters popularity amplification and outperforms previous debiasing methods on real-world datasets.
This paper proposes a Graph-Based Fraud Detection Model (DPF-GFD) with Dual-Path Graph Filtering to address challenges in fraud detection on graph data. It applies a beta wavelet-based operator and an improved low-pass filter on a similarity graph to capture key structural patterns.
This paper introduces Transductive Sharpening (TS), a novel loss-level modification for semi-supervised node classification. It leverages predictions on unlabeled nodes by minimizing prediction entropy to extract useful training signals often discarded by standard supervised objectives.
Mochi is a Graph Foundation Model that improves efficiency and task unification by employing a meta-learning based training framework. It pre-trains on few-shot episodes directly mirroring downstream evaluation, addressing limitations of traditional reconstruction-based pre-training and achieving competitive performance.
GraphDiffMed is a new framework for recommending safe and effective medication combinations from electronic health records (EHRs). It applies dual-scale Differential Attention to filter noisy signals and incorporates pharmacological constraints during learning, significantly improving recommendation quality.