Robust Link Prediction via Information Bottleneck-based Data Augmentation
CORE, a novel data augmentation framework, leverages the Information Bottleneck principle to eliminate noisy and spurious edges while recovering missing edges in graphs, thereby enhancing the generalizability of link prediction models.
VCR-Graphormer: A Mini-Batch Graph Transformer via Virtual Connections at ICLR 2024
Efficient mini-batch training with personalized PageRank tokenization in graph transformers.
Simplified PCNet with Robustness: Enhancing Graph Neural Networks for Real-World Data
Enhancing the robustness and efficiency of Simplified PCNet for graph representation learning.
Multi-Relational Graph Neural Network for Out-of-Domain Link Prediction
Novel Graph Neural Network GOOD designed for out-of-domain link prediction in dynamic multi-relational graphs.
Polynormer: Polynomial-Expressive Graph Transformer in Linear Time at ICLR 2024
Polynormer is a novel polynomial-expressive graph transformer with linear complexity that outperforms state-of-the-art GNN and GT baselines on various datasets.
Unveiling the Impact of Edge Perturbation on Graph Neural Networks
The author explores the dual effects of edge perturbation on graph neural networks, proposing a unified formulation to bridge augmentation and attack methods by prioritizing edge perturbations.
Revisiting Chebyshev Approximation for Spectral Graph Convolutions in Graph Neural Networks
ChebNet's inferior performance is due to illegal coefficients leading to over-fitting, addressed by ChebNetII using Chebyshev interpolation. ChebNetII outperforms SOTA methods in graph neural networks.
From Coupled Oscillators to Graph Neural Networks: Addressing Over-smoothing with Kuramoto Model-based Approach
The author proposes the KuramotoGNN, a continuous-depth GNN inspired by the Kuramoto model, to address over-smoothing by leveraging insights from synchronization. The model demonstrates superior performance compared to other GNN variants.
Polynormer: Polynomial-Expressive Graph Transformer in Linear Time
Polynormer introduces a polynomial-expressive graph transformer with linear complexity, balancing expressivity and scalability. The model outperforms state-of-the-art GNN and GT baselines on various datasets.
© 2024 by Linnk AI