Abstract:Graph Neural Networks (GNNs) have been broadly applied in many urban applications upon formulating a city as an urban graph whose nodes are urban objects like regions or points of interest. Recently, a few enhanced GNN architectures have been developed to tackle heterophily graphs where connected nodes are dissimilar. However, urban graphs usually can be observed to possess a unique spatial heterophily property; that is, the dissimilarity of neighbors at different spatial distances can exhibit great diversity. This property has not been explored, while it often exists. To this end, in this paper, we propose a metric, named Spatial Diversity Score, to quantitatively measure the spatial heterophily and show how it can influence the performance of GNNs. Indeed, our experimental investigation clearly shows that existing heterophilic GNNs are still deficient in handling the urban graph with high spatial diversity score. This, in turn, may degrade their effectiveness in urban applications. Along this line, we propose a Spatial Heterophily Aware Graph Neural Network (SHGNN), to tackle the spatial diversity of heterophily of urban graphs. Based on the key observation that spatially close neighbors on the urban graph present a more similar mode of difference to the central node, we first design a rotation-scaling spatial aggregation module, whose core idea is to properly group the spatially close neighbors and separately process each group with less diversity inside. Then, a heterophily-sensitive spatial interaction module is designed to adaptively capture the commonality and diverse dissimilarity in different spatial groups. Extensive experiments on three real-world urban datasets demonstrate the superiority of our SHGNN over several its competitors.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the problem of dealing with spatial heterophily in urban graphs. Specifically, the paper points out that existing graph neural networks (GNNs) have a serious limitation when dealing with urban graphs, that is, most of them are based on the assumption of homophily, that is, nodes in the graph tend to be connected to nodes with similar features or labels. However, in real - world urban graphs, the phenomenon of heterophily is widespread, that is, connected nodes have different features or labels. In particular, urban graphs usually exhibit unique spatial heterophily characteristics, that is, the differences between neighbors at different spatial distances show significant diversity. This characteristic has not been fully explored by existing methods, and it often has a negative impact on the effectiveness of urban applications. To meet this challenge, the paper proposes a new metric - the Spatial Diversity Score - to quantify the spatial heterophily of urban graphs and study its impact on the performance of existing GNNs. Experimental results show that existing heterophily GNNs perform poorly when dealing with urban graphs with high Spatial Diversity Scores, indicating the importance of developing new models that can effectively deal with diverse spatial heterophily in urban graphs. Therefore, the paper proposes a new graph neural network model - the Spatial Heterophily - Aware Graph Neural Network (SHGNN), which aims to deal with spatial heterophily in urban graphs through two specially designed modules: 1. **Rotation - Scaling Spatial Aggregation Module**: The core idea of this module is to reasonably group spatially close neighbors according to their positions relative to the central node and process the information within each group separately to reduce the diversity of intra - group heterophily. 2. **Heterophily - Sensitive Spatial Interaction Module**: This module designs two learnable kernel functions to adaptively capture the commonalities and diverse differences within different spatial groups, thereby determining the information required by the central node and its importance. Through these innovations, SHGNN can exhibit superior performance in dealing with spatial heterophily in urban graphs, especially on datasets with high Spatial Diversity Scores.

Spatial Heterophily Aware Graph Neural Networks

Beyond Homophily and Homogeneity Assumption: Relation-Based Frequency Adaptive Graph Neural Networks

Heterophilic Graph Neural Network Based on Spatial and Frequency Domain Adaptive Embedding Mechanism

Heterophilous Distribution Propagation for Graph Neural Networks

Addressing Heterogeneity and Heterophily in Graphs: A Heterogeneous Heterophilic Spectral Graph Neural Network

Revisiting Heterophily For Graph Neural Networks

Graph Neural Networks with Heterophily

Exploiting Neighbor Effect: Conv-Agnostic GNN Framework for Graphs with Heterophily

Learn from Heterophily: Heterophilous Information-enhanced Graph Neural Network

Graph Neural Networks for Graphs with Heterophily: A Survey

The Heterophilic Snowflake Hypothesis: Training and Empowering GNNs for Heterophilic Graphs

Seeking Similarities While Removing Differences: Graph Neural Networks Based on Node Correlation.

Finding Global Homophily in Graph Neural Networks When Meeting Heterophily.

Structure-Guided Input Graph for GNNs facing Heterophily

Scalable Graph Neural Networks for Heterogeneous Graphs

Quantifying Spatial Homogeneity of Urban Road Networks via Graph Neural Networks

SHGNN: Structure-Aware Heterogeneous Graph Neural Network

SR-HGN: Semantic- and Relation-Aware Heterogeneous Graph Neural Network

UD-GNN

HAGNN: Hybrid Aggregation for Heterogeneous Graph Neural Networks

Heterophily-Aware Graph Attention Network