Limits of Depth: Over-Smoothing and Over-Squashing in GNNs

Graph Neural Networks(GNNs)have become a widely used tool for learning and analyzing data on graph structures,largely due to their ability to preserve graph structure and properties via graph representation learning.However,the effect of depth on the performance of GNNs,particularly isotropic and anisotropic models,remains an active area of research.This study presents a comprehensive exploration of the impact of depth on GNNs,with a focus on the phenomena of over-smoothing and the bottleneck effect in deep graph neural networks.Our research investigates the tradeoff between depth and performance,revealing that increasing depth can lead to over-smoothing and a decrease in performance due to the bottleneck effect.We also examine the impact of node degrees on classification accuracy,finding that nodes with low degrees can pose challenges for accurate classification.Our experiments use several benchmark datasets and a range of evaluation metrics to compare isotropic and anisotropic GNNs of varying depths,also explore the scalability of these models.Our findings provide valuable insights into the design of deep GNNs and offer potential avenues for future research to improve their performance.

A binary-domain recurrent-like architecture-based dynamic graph neural network

The integration of Dynamic Graph Neural Networks(DGNNs)with Smart Manufacturing is crucial as it enables real-time,adaptive analysis of complex data,leading to enhanced predictive accuracy and operational efficiency in industrial environments.To address the problem of poor combination effect and low prediction accuracy of current dynamic graph neural networks in spatial and temporal domains,and over-smoothing caused by traditional graph neural networks,a dynamic graph prediction method based on spatiotemporal binary-domain recurrent-like architecture is proposed:Binary Domain Graph Neural Network(BDGNN).The proposed model begins by utilizing a modified Graph Convolutional Network(GCN)without an activation function to extract meaningful graph topology information,ensuring non-redundant embeddings.In the temporal domain,Recurrent Neural Network(RNN)and residual systems are employed to facilitate the transfer of dynamic graph node information between learner weights,aiming to mitigate the impact of noise within the graph sequence.In the spatial domain,the AdaBoost(Adaptive Boosting)algorithm is applied to replace the traditional approach of stacking layers in a graph neural network.This allows for the utilization of multiple independent graph learners,enabling the extraction of higher-order neighborhood information and alleviating the issue of over-smoothing.The efficacy of BDGNN is evaluated through a series of experiments,with performance metrics including Mean Average Precision(MAP)and Mean Reciprocal Rank(MRR)for link prediction tasks,as well as metrics for traffic speed regression tasks across diverse test sets.Compared with other models,the better experiments results demonstrate that BDGNN model can not only better integrate the connection between time and space information,but also extract higher-order neighbor information to alleviate the over-smoothing phenomenon of the original GCN.