115 Hits in 1.2 sec

Semi-Supervised Learning with Heterophily [article]

Wolfgang Gatterbauer
2016 arXiv   pre-print
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).  ...  graph-based label propagation algorithms by allowing them to propagate generalized assumptions about "attraction" or "compatibility" between classes of neighboring nodes (in particular those that involve heterophily  ...  Semi-Supervised Learning with Heterophily (SSL-H): We gen- eralize semi-supervised learning to general heterophily as- sumptions.  ... 
arXiv:1412.3100v2 fatcat:pjhrn5tmoveqtojact3a6itapi

Powerful Graph Convolutioal Networks with Adaptive Propagation Mechanism for Homophily and Heterophily [article]

Tao Wang and Rui Wang and Di Jin and Dongxiao He and Yuxiao Huang
2021 arXiv   pre-print
., 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).  ...  Existing methods deal with heterophily by mainly aggregating higher-order neighborhoods or combing the immediate representations, which leads to noise and irrelevant information in the result.  ...  In return, the propagation process can help learn better homophily degree matrix through downstream semi-supervised task.  ... 
arXiv:2112.13562v1 fatcat:jwi4mrv4t5hjll5die46e4duzy

Graph Neural Networks with Heterophily [article]

Jiong Zhu, Ryan A. Rossi, Anup Rao, Tung Mai, Nedim Lipka, Nesreen K. Ahmed, Danai Koutra
2021 arXiv   pre-print
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.  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Quadro P6000 GPU used for this research.  ... 
arXiv:2009.13566v3 fatcat:cwiivtjlxnck7bslu6fcvtx6oq

GBK-GNN: Gated Bi-Kernel Graph Neural Networks for Modeling Both Homophily and Heterophily [article]

Lun Du, Xiaozhou Shi, Qiang Fu, Hengyu Liu, Shi Han, Dongmei Zhang
2021 arXiv   pre-print
We conduct extensive experiments on various datasets with different homophily-heterophily properties.  ...  Graph Neural Networks (GNNs) are widely used on a variety of graph-based machine learning tasks.  ...  Node classification is a semi-supervised learning problem and it is formally defined as follows: Definition 1 (Node Classification).  ... 
arXiv:2110.15777v1 fatcat:2a2exdwqnjhmfmjykwczbdmcrq

Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs [article]

Jiong Zhu, Yujun Yan, Lingxiao Zhao, Mark Heimann, Leman Akoglu, Danai Koutra
2020 arXiv   pre-print
We investigate the representation power of graph neural networks in the semi-supervised node classification task under heterophily or low homophily, i.e., in networks where connected nodes may have different  ...  and real networks with heterophily, respectively, and yield competitive performance under homophily.  ...  We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Quadro P6000 GPU used for this research.  ... 
arXiv:2006.11468v2 fatcat:jcu5qvoffnca3jxkw7obrbcxuy

Simple Truncated SVD based Model for Node Classification on Heterophilic Graphs [article]

Vijay Lingam, Rahul Ragesh, Arun Iyer, Sundararajan Sellamanickam
2021 arXiv   pre-print
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.  ...  In the context of semi-supervised classification, this implies that nodes with similar labels are likely to be connected.  ...  Graph Neural Networks (GNNs) [5, 8, 14] leverage network information along with node features to improve their semi-supervised classification performance.  ... 
arXiv:2106.12807v1 fatcat:22kx74sgafdidm3b3cl2i63hca

Tree Decomposed Graph Neural Network [article]

Yu Wang, Tyler Derr
2021 arXiv   pre-print
smoothing between neighborhoods in different layers and can thus compromise the performance, especially on heterophily networks.  ...  Comprehensive experiments demonstrate the superior performance of TDGNN on both homophily and heterophily networks under a variety of node classification settings.  ...  Node Classification For the semi-supervised node classification task, we apply the fixed split following [38] and random training/validation/testing split on Cora, Citeseer, and Pubmed, with 20 nodes  ... 
arXiv:2108.11022v1 fatcat:5vnekq7t7zh2bkdrulg2zfgq6m

Simplifying approach to Node Classification in Graph Neural Networks [article]

Sunil Kumar Maurya, Xin Liu, Tsuyoshi Murata
2021 arXiv   pre-print
Through our experiments, we show that learning certain subsets of these features can lead to better performance on wide variety of datasets.  ...  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  ...  the conventional GNN formulation [1] using ̃ , a simple 2-layered GNN can be represented as, = ̃ ( ̃ (0) ) (1) (3) Node Classification Node classification is an extensively studied graph based semi-supervised  ... 
arXiv:2111.06748v1 fatcat:wswnvl3nonbmhi6jm2sn3cox3u

Effective Eigendecomposition based Graph Adaptation for Heterophilic Networks [article]

Vijay Lingam, Rahul Ragesh, Arun Iyer, Sundararajan Sellamanickam
2021 arXiv   pre-print
Performance improvement is achieved by learning flexible graph adaptation functions that modulate the eigenvalues of the graph.  ...  Several approaches address the issue of heterophily by proposing models that adapt the graph by optimizing task-specific loss function using labelled data.  ...  In the context of semi-supervised classification, this implies that nodes with similar labels are likely to be connected.  ... 
arXiv:2107.13312v1 fatcat:dcn2oklm3ze5lmp5nrrqhohruq

FDGATII : Fast Dynamic Graph Attention with Initial Residual and Identity Mapping [article]

Gayan K. Kulatilleke, Marius Portmann, Ryan Ko, Shekhar S. Chandra
2021 arXiv   pre-print
We show that FDGATII outperforms GAT and GCN based benchmarks in accuracy and performance on fully supervised tasks, obtaining state-of-the-art results on Chameleon and Cornell datasets with zero domain-specific  ...  While Graph Neural Networks have gained popularity in multiple domains, graph-structured input remains a major challenge due to (a) over-smoothing, (b) noisy neighbours (heterophily), and (c) the suspended  ...  The first property allows us to put strong regularization on W l to avoid over-fitting, while the later is desirable in semi-supervised tasks where training data is limited.  ... 
arXiv:2110.11464v2 fatcat:bddix5udzvfujddf6kcri32v4i

On the Relationship between Heterophily and Robustness of Graph Neural Networks [article]

Jiong Zhu, Junchen Jin, Donald Loveland, Michael T. Schaub, Danai Koutra
2021 arXiv   pre-print
We theoretically and empirically show that for graphs exhibiting homophily (low heterophily), impactful structural attacks always lead to increased levels of heterophily, while for graph with heterophily  ...  Furthermore, models with this design can be readily combined with explicit defense mechanisms to yield improved robustness with up to 18.33% increase in performance under attacks compared to the best-performing  ...  We gratefully acknowledge the support of NVIDIA Corporation with  ... 
arXiv:2106.07767v2 fatcat:y2j3ox4ibbar3bl6coiesry32a

Is Homophily a Necessity for Graph Neural Networks? [article]

Yao Ma, Xiaorui Liu, Neil Shah, Jiliang Tang
2021 arXiv   pre-print
When applied to semi-supervised node classification, GNNs are widely believed to work well due to the homophily assumption ("like attracts like"), and fail to generalize to heterophilous graphs where dissimilar  ...  Graph neural networks (GNNs) have shown great prowess in learning representations suitable for numerous graph-based machine learning tasks.  ...  The goal of semi-supervised node classification (SSNC) is to learn a mapping f : V → C utilizing the graph G, the node features X and the labels for nodes in V label .  ... 
arXiv:2106.06134v3 fatcat:wj5b47neorfpjjm3sulgydrqye

Graph Pointer Neural Networks [article]

Tianmeng Yang, Yujing Wang, Zhihan Yue, Yaming Yang, Yunhai Tong, Jing Bai
2022 arXiv   pre-print
Most existing GNNs assume strong homophily of graph structure and apply permutation-invariant local aggregation of neighbors to learn a representation for each node.  ...  The pointer-network-based ranker in GPNN is joint-optimized with other parts in an end-to-end manner.  ...  Here we focus on a brief introduction of representation learning for graph nodes in a supervised or semi-supervised setting.  ... 
arXiv:2110.00973v2 fatcat:el2htkjkc5d33jsjnmneb2o5mq

An Interpretable Graph Generative Model with Heterophily [article]

Sudhanshu Chanpuriya, Ryan A. Rossi, Anup Rao, Tung Mai, Nedim Lipka, Zhao Song, Cameron Musco
2021 arXiv   pre-print
to capture both heterophily and homophily in the data.  ...  These models output the probabilities of edges existing between all pairs of nodes, and the probability of a link between two nodes increases with the dot product of vectors associated with the nodes.  ...  bound to the intensities of community participation (i.e. the entries of each f ), so it is unclear how to incorporate prior knowledge about community membership in the form of binary labels, as in a semi-supervised  ... 
arXiv:2111.03030v1 fatcat:sc64kdqmnzdtvabhtajw6wqkum

Label Propagation on K-partite Graphs with Heterophily [article]

Dingxiong Deng, Fan Bai, Yiqi Tang, Shuigeng Zhou, Cyrus Shahabi, Linhong Zhu
2017 arXiv   pre-print
With this model, we develop a novel label inference algorithm framework with update rules in near-linear time complexity.  ...  In this paper, for the first time, we study label propagation in heterogeneous graphs under heterophily assumption.  ...  with fewer labeled data), but rather to propose a better semi-supervised label propagation algorithm for tripartite graphs.  ... 
arXiv:1701.06075v1 fatcat:4cvenwhqpbcgxkkwe7g6cbcl7y
« Previous Showing results 1 — 15 out of 115 results