A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Filters
An Interpretable Graph Generative Model with Heterophily
[article]
2021
arXiv
pre-print
Preserved Fulltext
We propose the first edge-independent graph generative model that is a) expressive enough to capture heterophily, b) produces nonnegative embeddings, which allow link predictions to be interpreted in terms ...
Many models for graphs fall under the framework of edge-independent dot product models. ...
CONCLUSION We introduce an interpretable, edge-independent graph generative model that is highly expressive at representing both heterophily and overlapping communities. ...
arXiv:2111.03030v1
fatcat:sc64kdqmnzdtvabhtajw6wqkum
Graph Neural Networks with Heterophily
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Other Versions
The proposed framework incorporates an interpretable compatibility matrix for modeling the heterophily or homophily level in the graph, which can be learned in an end-to-end fashion, enabling it to go ...
In this work, we propose a novel framework called CPGNN that generalizes GNNs for graphs with either homophily or heterophily. ...
W911NF1810397, an Adobe Digital Experience research faculty award, an Amazon faculty award, a Google faculty award, and AWS Cloud Credits for Research. ...
arXiv:2009.13566v3
fatcat:cwiivtjlxnck7bslu6fcvtx6oq
Graph Neural Networks for Graphs with Heterophily: A Survey
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Specifically, we propose a systematic taxonomy that essentially governs existing heterophilic GNN models, along with a general summary and detailed analysis. ...
However, as a ubiquitous graph property in numerous real-world scenarios, heterophily, i.e., nodes with different labels tend to be linked, significantly limits the performance of tailor-made homophilic ...
Currently, most of heterophilic GNNs are generally evaluated on the first 6 benchmark datasets in Table 2 , with node heterophily and edge heterophily to measure the heterophilic degree of a graph. ...
arXiv:2202.07082v1
fatcat:26d5mpnfzbb7fmloytdbj67yxa
GBK-GNN: Gated Bi-Kernel Graph Neural Networks for Modeling Both Homophily and Heterophily
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We conduct extensive experiments on various datasets with different homophily-heterophily properties. ...
For node-level tasks, GNNs have strong power to model the homophily property of graphs (i.e., connected nodes are more similar) while their ability to capture the heterophily property is often doubtful ...
GENERALIZATION BOUND OF GCN WITH REGARD TO HOMOPHILY RATIO In this section, we theoretically analyze the generalization ability of Graph Convolutional Networks (GCNs) with regard to different homophily-heterophily ...
arXiv:2110.15777v3
fatcat:vy3fql3avjbavoq3bkcoucg7ba
Demystifying Graph Convolution with a Simple Concatenation
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
in the heterophily case. ...
On the other hand, many GConv-based models do not quantify the effect of graph topology and node features on performance, and are even surpassed by some models that do not consider graph structure or node ...
For graph concatenation model, we adopt Deepwalk [20] to generate graph embeddings which are later concatenated with the node features. ...
arXiv:2207.12931v1
fatcat:corlluporbhkdgtd6upj4z5lt4
Unifying Homophily and Heterophily Network Transformation via Motifs
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
H2NT can be used as an enhancer to be integrated with any existing network embedding methods without requiring any changes to latter methods. ...
However, there is no general and universal solution that takes both into consideration. ...
A recent transformation model graph diffusion convolution (GDC) [Klicpera et al., 2019] generates a new network by constructing a diffusion graph obtained by a polynomial function, and then sparsify ...
arXiv:2012.11400v2
fatcat:llwy5qsp6beuhf57d5gojadxbm
NCGNN: Node-Level Capsule Graph Neural Network for Semisupervised Classification
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
In this paper, we propose a Node-level Capsule Graph Neural Network (NCGNN) to address these problems with an improved message passing scheme. ...
Therefore, it can relieve the over-smoothing issue and learn effective node representations over graphs with homophily or heterophily. ...
viewed as an ensemble of GNN models. ...
arXiv:2012.03476v2
fatcat:cuy5bd5ldbcbzfbkkwqpcwewt4
Powerful Graph Convolutioal Networks with Adaptive Propagation Mechanism for Homophily and Heterophily
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
., nodes with same class are prone to connect to each other), while ignoring the heterophily which exists in many real-world networks (i.e., nodes with different classes tend to form edges). ...
Then we incorporate the learnable homophily degree into the graph convolution framework, which is trained in an end-to-end schema, enabling it to go beyond the assumption of homophily. ...
CPGNN (Zhu et al. 2021 ) incorporates an interpretable compatibility matrix for modeling the heterophily or homophily level in graphs, enabling it to go beyond the assumption of strong homophily. ...
arXiv:2112.13562v1
fatcat:jwi4mrv4t5hjll5die46e4duzy
Exploiting Neighbor Effect: Conv-Agnostic GNNs Framework for Graphs with Heterophily
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Due to the homophily assumption in graph convolution networks, a common consensus is that graph neural networks (GNNs) perform well on homophilic graphs but may fail on heterophilic graphs with many inter-class ...
Extensive ablation studies and robustness analysis further verify the effectiveness, robustness, and interpretability of our framework. ...
GPRGNN [13] modifies the convolution to the generalized page rank and learns an arbitrary K-order polynomial graph filter. ...
arXiv:2203.11200v2
fatcat:g3i3qyabavhbtflp3xpbi4htry
Simple Truncated SVD based Model for Node Classification on Heterophilic Graphs
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Graph Neural Networks (GNNs) have shown excellent performance on graphs that exhibit strong homophily with respect to the node labels i.e. connected nodes have same labels. ...
This work is an early investigation into methods that differ from aggregation based approaches. ...
GPR-GNN [4] takes the idea proposed in APPNP and generalizes the Pagerank model that works well for graphs with varying homophily scores. ...
arXiv:2106.12807v1
fatcat:22kx74sgafdidm3b3cl2i63hca
Semi-Supervised Learning with Heterophily
[article]
2016
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We thus call this formulation Semi-Supervised Learning with Heterophily (SSLH) and show how it generalizes and improves upon a recently proposed approach called Linearized Belief Propagation (LinBP). ...
We derive a family of linear inference algorithms that generalize existing graph-based label propagation algorithms by allowing them to propagate generalized assumptions about "attraction" or "compatibility ...
I would like to thank Christos Faloutsos for continued support and Stephan Günneman for insightful comments on an early version of this paper. ...
arXiv:1412.3100v2
fatcat:pjhrn5tmoveqtojact3a6itapi
Improving Graph Neural Networks with Simple Architecture Design
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Finally, we demonstrate with experiments that the model is scalable for large graphs with millions of nodes and billions of edges. ...
These graphs are often created with assumed intrinsic relations between the entities. ...
PRELIMINARIES Let = ( , ) be an undirected graph with nodes and edges. ...
arXiv:2105.07634v1
fatcat:q4lxkvn2ffa75cwox24lrqndte
How does Heterophily Impact Robustness of Graph Neural Networks? Theoretical Connections and Practical Implications
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
Additionally, combining this design with explicit defense mechanisms against adversarial attacks leads to an improved robustness with up to 18.33% performance increase under attacks compared to the best-performing ...
vaccinated model. ...
W911NF1810397, an Adobe Digital Experience research faculty award, an Amazon faculty award, a Google faculty award, and AWS Cloud Credits for Research. ...
arXiv:2106.07767v3
fatcat:conprpfs6fe4ld6yprlikzyho4
Simplifying approach to Node Classification in Graph Neural Networks
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
than state-of-the-art GNN models in nine benchmark datasets for the node classification task, with remarkable improvements up to 51.1%. ...
As different graph datasets show varying levels of homophily and heterophily in features and class label distribution, it becomes essential to understand which features are important for the prediction ...
Preliminaries Let = ( , ) be an undirected graph with nodes and edges. ...
arXiv:2111.06748v1
fatcat:wswnvl3nonbmhi6jm2sn3cox3u
Label-informed Graph Structure Learning for Node Classification
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
To tackle this problem, some studies exploit different graph structure learning strategies to refine the original graph structure. ...
Graph Neural Networks (GNNs) have achieved great success among various domains. Nevertheless, most GNN methods are sensitive to the quality of graph structures. ...
We train the model using Adam optimizer [5] with an initial learning rate of 0.01. ...
arXiv:2108.04595v1
fatcat:l5f5fjf5obajhfkej6sgiizjui
« Previous
Showing results 1 — 15 out of 526 results