10,031 Hits in 6.4 sec

Classic Graph Structural Features Outperform Factorization-Based Graph Embedding Methods on Community Labeling [article]

Andrew Stolman, Caleb Levy, C. Seshadhri, Aneesh Sharma
In constrast, a simple logistic model with classic graph structural features handily outperforms the embedding models.  ...  In all cases we studied, the models trained from embedding features perform poorly on community labeling.  ...  Evaluations on real data: In our experiments, not only do we see poor absolute performance on the community labeling task, but a baseline LR-Structural based on "classic" graph features handily outperforms  ... 
doi:10.48550/arxiv.2201.08481 fatcat:eggyihzfknasfjz724rzaiwswi

Neighbor2vec: an efficient and effective method for Graph Embedding [article]

Zhiming Lin
2022 arXiv   pre-print
The neighbor2vec's representations are able to outperform all baseline methods and two classical GNN models in all six experiments.  ...  This paper propose neighbor2vec, a neighbor-based sampling strategy used algorithm to learn the neighborhood representations of node, a framework to gather the structure information by feature propagation  ...  By using stochastic gradient descent based matrix factorization, it finds a way to get the embedding of a network. The object of graph factorization, however, is not suitable for network.  ... 
arXiv:2201.02626v1 fatcat:fb7ekwgc3nd4vihyiymjcuuhza

Learning Universal Network Representation via Link Prediction by Graph Convolutional Neural Network

Weiwei Gu, Fei Gao, Ruiqi Li, Jiang Zhang
2021 Journal of Social Computing  
Experiments prove the effectiveness of LPNR on three real-world networks. With the mini-batch and fixed sampling strategy, LPNR can learn the embedding of large graphs in a few hours.  ...  In this paper, we propose a novel network representation method, named Link Prediction based Network Representation (LPNR), which generalizes the latest graph neural network and optimizes a carefully designed  ...  Over the past decades, several embedding algorithms have been proposed and they can be classified into the following categories: matrix factorization based methods, random walk based methods, graph neural  ... 
doi:10.23919/jsc.2021.0001 fatcat:q4vwqeuudbglzdykpwdhst2mwa

ACE: Ant Colony based Multi-level Network Embedding for Hierarchical Graph Representation Learning

Jianming Lv, Jiajie Zhong, Jintao Liang, Zhenguo Yang
2019 IEEE Access  
To solve this problem, we propose ACE, a novel network embedding method, to preserve the features of hierarchical clustering structures.  ...  Then, we generate the embedding vectors from multiple layers of the graph pyramid and blend these multi-level vectors into the final representation of nodes based on the PCA dimension reduction algorithm  ...  the Label Propagation Algorithm (LPA) [20] , [27] on a graph and merges the nodes in one community.  ... 
doi:10.1109/access.2019.2920671 fatcat:5xnmgbhacraf3jyxxc5p4vk5mu

GrAMME: Semi-Supervised Learning using Multi-layered Graph Attention Models [article]

Uday Shankar Shanthamallu, Jayaraman J. Thiagarajan, Huan Song and Andreas Spanias
2019 arXiv   pre-print
DeepWalk, are widely adopted for community discovery, we argue that feature learning with random node attributes, using graph neural networks, can be more effective.  ...  multi-layered graph embeddings.  ...  Prior work on multi-layered graphs focuses extensively on unsupervised community detection, and they can be broadly classified into methods that obtain a consensus community structure for producing node  ... 
arXiv:1810.01405v2 fatcat:gydvjhal4jfbfi24bvnbsrpr2u


Anton Tsitsulin, Davide Mottin, Panagiotis Karras, Emmanuel Müller
2018 Proceedings of the 2018 World Wide Web Conference on World Wide Web - WWW '18  
Our experimental study on standard benchmarks and real-world datasets demonstrates that VERSE, instantiated with diverse similarity measures, outperforms state-of-the-art methods in terms of precision  ...  Past research has addressed the problem of extracting such embeddings by adopting methods from words to graphs, without defining a clearly comprehensible graph-related objective.  ...  Graph Reconstruction Good graph embeddings should preserve the graph structure in the embedding space. We evaluate the performance of our method on reconstructing the graph's adjacency matrix.  ... 
doi:10.1145/3178876.3186120 dblp:conf/www/TsitsulinMKM18 fatcat:gn6asmflqfffzfmqmf72xmn7wy


Benedek Rozemberczki, Ryan Davies, Rik Sarkar, Charles Sutton
2019 Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining  
GEMSEC is a general extension of earlier work in the domain of sequence-based graph embedding.  ...  Modern graph embedding procedures can efficiently process graphs with millions of nodes.  ...  Classical approaches to community detection depend on properties such as graph metrics, spectral properties and density of shortest paths [4] .  ... 
doi:10.1145/3341161.3342890 dblp:conf/asunam/RozemberczkiDSS19 fatcat:ewiv4ic67ffr5mzyjpvre5anxm

Simplification of Graph Convolutional Networks: A Matrix Factorization-based Perspective [article]

Qiang Liu and Haoli Zhang and Zhaocheng Liu
2020 arXiv   pre-print
Moreover, we have also conducted experiments on a typical task of graph embedding, i.e., community detection, and the proposed UCMF model outperforms several representative graph embedding models.  ...  Fortunately, methods based on Matrix Factorization (MF) naturally support constructing mini-batches, and thus are more friendly to distributed computing compared with GCN.  ...  Therefore, GCN is consistent with graph embedding methods in capturing the structural information.  ... 
arXiv:2007.09036v5 fatcat:vsljswaxnjdx7jz5uokye3y6pu

ExEm: Expert Embedding using dominating set theory with deep learning approaches [article]

N. Nikzad-Khasmakhi, M. A. Balafar, M.Reza Feizi-Derakhshi, Cina Motamed
2021 arXiv   pre-print
In this paper, we propose a graph embedding method, called ExEm, that uses dominating-set theory and deep learning approaches to capture node representations.  ...  To perform the analysis, graph embedding techniques have emerged as an effective and promising tool. Graph embedding attempts to represent graph nodes as low-dimensional vectors.  ...  On the other hand, a matrix factorization based method represents the connections between nodes as a matrix and factorizes this matrix to extract node embedding.  ... 
arXiv:2001.08503v2 fatcat:zolbpnm5draixbt3zbfy7hgena

Inductive Graph Embeddings through Locality Encodings [article]

Nurudin Alvarez-Gonzalez, Andreas Kaltenbrunner, Vicenç Gómez
2020 arXiv   pre-print
Despite the overwhelming number of existing methods, is is unclear how to exploit network structure in a way that generalizes easily to unseen nodes, edges or graphs.  ...  tasks, when used as additional features in a neural network.  ...  tasks. • The proposed local encodings generalize to unseen graphs from the same domain in a multi-label node classification setting, outperforming methods based on Graph Convolutional Networks. • Our  ... 
arXiv:2009.12585v1 fatcat:wbpevwstpnd33l3yssdqgicjlu

Exponential Family Graph Embeddings [article]

Abdulkadir Çelikkanat, Fragkiskos D. Malliaros
2019 arXiv   pre-print
Our experimental evaluation on real-world datasets demonstrates that the proposed techniques outperform well-known baseline methods in two downstream machine learning tasks.  ...  We introduce the generic exponential family graph embedding model, that generalizes random walk-based network representation learning techniques to exponential family conditional distributions.  ...  the proposed exponential family graph embedding models generally outperform widely used baseline approaches in various learning tasks on graphs.  ... 
arXiv:1911.09007v1 fatcat:lgwiuh3zuffznb7rdsnf2cghre

Correlation between the Dissemination of Classic English Literary Works and Cultural Cognition in the New Media Era

Weiwei Guo, Qiangyi Li
2022 Advances in Multimedia  
To get the best sequence label and get the last labeled entity information, in terms of knowledge graph construction and visual query, a workflow method for building knowledge graph from unstructured text  ...  Based on n-gram embedding, by combining pretraining embedding and radical embedding, it can fully consider the rich semantic information in English literature works to extract.  ...  In particular, since the radical embedding uses the CNN network structure, using it to extract image features consumes more GPU resources.  ... 
doi:10.1155/2022/3616432 fatcat:iyhwwr2sgbhmlna7nwbdjx6up4

Exponential Family Graph Embeddings

Abdulkadir Celikkanat, Fragkiskos D. Malliaros
Our experimental evaluation on real-world datasets demonstrates that the proposed techniques outperform well-known baseline methods in two downstream machine learning tasks.  ...  We introduce the generic exponential family graph embedding model, that generalizes random walk-based network representation learning techniques to exponential family conditional distributions.  ...  The authors of (Wang et al. 2017) , proposed a matrix factorization algorithm that incorporates the community structure into the embedding process, implicitly focusing on the quantity of modularity.  ... 
doi:10.1609/aaai.v34i04.5737 fatcat:ldemjmb4kbhh7ed5vnfshrri3e

Computational Drug-target Interaction Prediction based on Graph Embedding and Graph Mining

Maha A. Thafar, Somayah Albaradie, Rawan S. Olayan, Haitham Ashoor, Magbubah Essack, Vladimir B. Bajic
2020 Proceedings of the 2020 10th International Conference on Bioscience, Biochemistry and Bioinformatics  
In this work, we propose a computational method for (Drug-Target interaction prediction using Graph Embedding and graph Mining), DTiGEM.  ...  Specifically, we demonstrate that based on the average AUPR score across all benchmark datasets, DTiGEM achieves the highest average AUPR value (0.831), thus reducing the prediction error by 22.4% relative  ...  DTiGEM combines similarity-based as well as feature-based techniques. It uses graph embedding, graph-mining, and ML.  ... 
doi:10.1145/3386052.3386062 fatcat:o2hk74kai5gfbhiouaclqcltj4

Going Deep: Graph Convolutional Ladder-Shape Networks

Ruiqi Hu, Shirui Pan, Guodong Long, Qinghua Lu, Liming Zhu, Jing Jiang
The proposed GCLN outperforms original GCNs, deep GCNs and other state-of-the-art GCN-based models for all three tasks, which were designed from various perspectives on six real-world benchmark data sets  ...  Neighborhood aggregation algorithms like spectral graph convolutional networks (GCNs) formulate graph convolutions as a symmetric Laplacian smoothing operation to aggregate the feature information of one  ...  Each given graph is embedded into a low-dimensional vector based on which the class label of the whole graph can be inferred.  ... 
doi:10.1609/aaai.v34i03.5673 fatcat:iqkayosfjbeijhy4zkyolhwyxi
« Previous Showing results 1 — 15 out of 10,031 results