Merit: multi-level graph embedding refinement framework for large-scale graph
Abstract The development of the Internet and big data has led to the emergence of graphs as an important data representation structure in various real-world scenarios. However, as data size increases, computational complexity and memory requirements pose significant challenges for graph embedding. T...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2023-08-01
|
| Series: | Complex & Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s40747-023-01211-3 |
| _version_ | 1797272232774336512 |
|---|---|
| author | Weishuai Che Zhaowei Liu Yingjie Wang Jinglei Liu |
| author_facet | Weishuai Che Zhaowei Liu Yingjie Wang Jinglei Liu |
| author_sort | Weishuai Che |
| collection | DOAJ |
| description | Abstract The development of the Internet and big data has led to the emergence of graphs as an important data representation structure in various real-world scenarios. However, as data size increases, computational complexity and memory requirements pose significant challenges for graph embedding. To address this challenge, this paper proposes a multilevel embedding refinement framework (MERIT) based on large-scale graphs, using spectral distance-constrained graph coarsening algorithms and an improved graph convolutional neural network model that addresses the over-smoothing problem by incorporating initial values and identity mapping. Experimental results on large-scale datasets demonstrate the effectiveness of MERIT, with an average AUROC score 8% higher than other baseline methods. Moreover, in a node classification task on a large-scale graph with 126,825 nodes and 22,412,658 edges, the framework improves embedding quality while enhancing the runtime by 25 times. The experimental findings highlight the superior efficiency and accuracy of the proposed approach compared to other graph embedding methods. |
| first_indexed | 2024-03-07T14:25:22Z |
| format | Article |
| id | doaj.art-945a391d019049618f2df45390fbda2b |
| institution | Directory Open Access Journal |
| issn | 2199-4536 2198-6053 |
| language | English |
| last_indexed | 2024-03-07T14:25:22Z |
| publishDate | 2023-08-01 |
| publisher | Springer |
| record_format | Article |
| series | Complex & Intelligent Systems |
| spelling | doaj.art-945a391d019049618f2df45390fbda2b2024-03-06T08:07:20ZengSpringerComplex & Intelligent Systems2199-45362198-60532023-08-011011303131810.1007/s40747-023-01211-3Merit: multi-level graph embedding refinement framework for large-scale graphWeishuai Che0Zhaowei Liu1Yingjie Wang2Jinglei Liu3School of Computer and Control Engineering, Yantai UniversitySchool of Computer and Control Engineering, Yantai UniversitySchool of Computer and Control Engineering, Yantai UniversitySchool of Computer and Control Engineering, Yantai UniversityAbstract The development of the Internet and big data has led to the emergence of graphs as an important data representation structure in various real-world scenarios. However, as data size increases, computational complexity and memory requirements pose significant challenges for graph embedding. To address this challenge, this paper proposes a multilevel embedding refinement framework (MERIT) based on large-scale graphs, using spectral distance-constrained graph coarsening algorithms and an improved graph convolutional neural network model that addresses the over-smoothing problem by incorporating initial values and identity mapping. Experimental results on large-scale datasets demonstrate the effectiveness of MERIT, with an average AUROC score 8% higher than other baseline methods. Moreover, in a node classification task on a large-scale graph with 126,825 nodes and 22,412,658 edges, the framework improves embedding quality while enhancing the runtime by 25 times. The experimental findings highlight the superior efficiency and accuracy of the proposed approach compared to other graph embedding methods.https://doi.org/10.1007/s40747-023-01211-3Graph representation learningGraph embeddingGraph neural networksGraph convolutional networkLarge-scale graph |
| spellingShingle | Weishuai Che Zhaowei Liu Yingjie Wang Jinglei Liu Merit: multi-level graph embedding refinement framework for large-scale graph Complex & Intelligent Systems Graph representation learning Graph embedding Graph neural networks Graph convolutional network Large-scale graph |
| title | Merit: multi-level graph embedding refinement framework for large-scale graph |
| title_full | Merit: multi-level graph embedding refinement framework for large-scale graph |
| title_fullStr | Merit: multi-level graph embedding refinement framework for large-scale graph |
| title_full_unstemmed | Merit: multi-level graph embedding refinement framework for large-scale graph |
| title_short | Merit: multi-level graph embedding refinement framework for large-scale graph |
| title_sort | merit multi level graph embedding refinement framework for large scale graph |
| topic | Graph representation learning Graph embedding Graph neural networks Graph convolutional network Large-scale graph |
| url | https://doi.org/10.1007/s40747-023-01211-3 |
| work_keys_str_mv | AT weishuaiche meritmultilevelgraphembeddingrefinementframeworkforlargescalegraph AT zhaoweiliu meritmultilevelgraphembeddingrefinementframeworkforlargescalegraph AT yingjiewang meritmultilevelgraphembeddingrefinementframeworkforlargescalegraph AT jingleiliu meritmultilevelgraphembeddingrefinementframeworkforlargescalegraph |