Node Embedding Preserving Graph Summarization
Abstract
1 Introduction
2 Related Work
2.1 Graph Summarization
2.2 Graph Summarization Preserving Node Embeddings
3 CR Reconstruction Scheme
Symbol | Definition |
---|---|
\(\mathcal {G}\) = \((\mathcal {V}, \mathcal {E})\) | Original graph with nodeset \(\mathcal {V}\) and edgeset \(\mathcal {E}\) |
\(\mathcal {G}_s\) = \((\mathcal {V}_s, \mathcal {E}_s)\) | Summary graph with supernodes \(\mathcal {V}_s\) and superedges \(\mathcal {E}_s\) |
\(\mathcal {G}_r\) = \((\mathcal {V}, \mathcal {E}_r)\) | Reconstructed graph with nodeset \(\mathcal {V}\) and edgeset \(\mathcal {E}_r\) |
\(v_i\) | Node i in the original graph \(\mathcal {G}\) |
\(\mathcal {S}_k\) | Supernode k in the summary graph \(\mathcal {G}_s\) |
\(d_i, D_k\) | Degree of node i and supernode k |
\(\mathbf {A}, \mathbf {A}_s, \mathbf {A}_r\) | Adjacency matrix of original, summary, reconstructed graphs |
\(\mathbf {D}, \mathbf {D}_s\) | Degree matrix of original and summary graphs |
\(\mathbf {L}, \mathbf {L}_s, \mathbf {L}_r\) | (Combinatorial) Laplacian matrices of original, summary, reconstructed graphs |
\(\mathbf {\mathcal {A}}, \mathbf {\mathcal {L}}\) | Normalized adjacency matrix and normalized Laplacian matrix |
\(\mathbf {P}, \mathbf {Q}\) | Membership and reconstruction matrix in summarization |
\(\mathbf {R}\) | Restoration matrix for recovering the original embeddings |
\(\mathbf {E}, \mathbf {E}_{s}\) | Embeddings of original graph and summary graph |
3.1 Graph Summarization and Reconstruction Scheme
4 Connection with Node Embedding Methods
4.1 Matrix-factorization-based Node Embedding Methods
4.2 Approximating Kernel Matrix
4.3 Approximating Node Embeddings
5 Proposed Methods
5.1 Kernel Matrix Error Analysis
5.2 HCSumm
6 Experiments
6.1 Experimental Setup
Dataset | #Nodes | #Edges | #Labels |
---|---|---|---|
Cora | 2,307 | 5,278 | 7 |
BlogCatalog | 10,312 | 667,966 | 39 |
Flickr | 89,250 | 5,899,882 | 195 |
YouTube | 1,138,499 | 2,990,443 | 47 |
6.2 Summary Quality
6.3 Node Embedding Preservation
6.3.1 NetMF.
6.3.2 DeepWalk.
6.4 Scalability
7 Conclusion
Footnotes
Appendix
A Proofs
A.1 Proof of Theorem 1
References
Index Terms
- Node Embedding Preserving Graph Summarization
Recommendations
Lossless graph summarization using dense subgraphs discovery
IMCOM '15: Proceedings of the 9th International Conference on Ubiquitous Information Management and CommunicationDense subgraph discovery, in a large graph, is useful to solve the community search problem. Motivated from this, we propose a graph summarization method where we search and aggregate dense subgraphs into super nodes. Since the dense subgraphs have high ...
PartKG2Vec: Embedding of Partitioned Knowledge Graphs
Knowledge Science, Engineering and ManagementAbstractLarge-scale knowledge graphs with billions of nodes and edges are increasingly common in many domains. Such graphs often exceed the capacity of the systems storing the graphs in a centralized data store, not to mention the limits of today’s graph ...
Hierarchical graph embedding in vector space by graph pyramid
Loss of information is the major challenge in graph embedding in vector space which reduces the impact of representational power of graphs in pattern recognition tasks. The objective of this article is to present a hierarchical framework which can ...
Comments
Information & Contributors
Information
Published In
Publisher
Association for Computing Machinery
New York, NY, United States
Publication History
Check for updates
Author Tags
Qualifiers
- Research-article
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 474Total Downloads
- Downloads (Last 12 months)474
- Downloads (Last 6 weeks)116
Other Metrics
Citations
View Options
Get Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Sign in