Community-aware dynamic network embedding by using deep autoencoder

Ma, Lijia; Zhang, Yutao; Li, Jianqiang; Lin, Qiuzhen; Bao, Qing; Wang, Shanfeng; Gong, Maoguo

doi:10.1016/j.ins.2020.01.027

Cited by 13 publications

(6 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Three types of downstream tasks are employed to evaluate the quality of node embeddings. Graph Reconstruction (GR) tasks [10], [16], [21] use current embeddings to reconstruct snapshot (by retrieving node neighbors for every node) at current timestep. Node Recommendation (NR) tasks [27] use current embeddings to recommend node neighbors for those affected nodes (by verifying their old and new neighbors) at next timestep.…”

Section: Discussionmentioning

confidence: 99%

“…The reasons of using it rather than other embedding approaches are as follows. While extending SNE to DNE approaches, first, the matrix factorization based and auto-encoder based approaches often require to set a fixed number of nodes in advance [9], [10], [16], [17], and hence are hard to handle unlimited new nodes. Second, the performance of graph convolution based approach [36], [37] largely depends on the node attribute, which is out of the consideration of this work.…”

Section: B Methods Descriptionmentioning

confidence: 99%

“…[14] proposes a heuristic strategy to adjust the Auto-Encoder to handle the challenge. [12], [15], [16] also adopt an Auto-Encoder to incrementally learn node embeddings. But these works do not adjust the architecture of the Auto-Encoder and hence, they need to fix the node size in advance.…”

Section: B Related Workmentioning

confidence: 99%

“…However, whether the transferred historical knowledge is beneficial to current model would depend on the dynamic nature (i.e., future changes between consecutive snapshots) of a network. Regarding existing DNE works, many of them such as [3], [9]- [11], [16], [18], [27], [38] assume that the changes between consecutive snapshots are smooth or small. It seems that the smoothness of changes is an important factor for DNE methods, especially for those DNE methods following the incremental learning paradigm.…”

Section: Incremental Learning and Robustness Of Dnementioning

confidence: 99%

See 3 more Smart Citations

Towards Robust Dynamic Network Embedding

Hou

Tang

2021

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence

View full text Add to dashboard Cite

Dynamic Network Embedding (DNE) has recently drawn much attention due to the dynamic nature of many real-world networks. Comparing to a static network, a dynamic network has a unique character called the degree of changes, which can be defined as the average number of the changed edges between consecutive snapshots spanning a dynamic network. The degree of changes could be quite different even for the dynamic networks generated from the same dataset. It is natural to ask whether existing DNE methods are effective and robust w.r.t. the degree of changes. Towards robust DNE, we suggest two important scenarios. One is to investigate the robustness w.r.t. different slicing settings that are used to generate different dynamic networks with different degree of changes, while another focuses more on the robustness w.r.t. different number of changed edges over timesteps.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: B Methods Descriptionmentioning

confidence: 99%

Section: B Related Workmentioning

confidence: 99%

Section: Incremental Learning and Robustness Of Dnementioning

confidence: 99%

See 2 more Smart Citations

Towards Robust Dynamic Network Embedding

Hou

Tang

2021

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…To solve this problem, Zhang et al [ 27 ] proposed a temporal deep autoencoder architecture that considered the graph structure and vertex attributes to test the community. Ma et al [ 28 ] proposed a community-aware dynamic network embedding method based on an autoencoder to record the dynamics of community structures. The results showed the proposed model performed well on existing graph issues (i.e., link prediction, network reconstruction).…”

Section: Related Workmentioning

confidence: 99%

Dynamic graph embedding for outlier detection on multiple meteorological time series

Jung

2021

PLoS ONE

View full text Add to dashboard Cite

Existing dynamic graph embedding-based outlier detection methods mainly focus on the evolution of graphs and ignore the similarities among them. To overcome this limitation for the effective detection of abnormal climatic events from meteorological time series, we proposed a dynamic graph embedding model based on graph proximity, called DynGPE. Climatic events are represented as a graph where each vertex indicates meteorological data and each edge indicates a spurious relationship between two meteorological time series that are not causally related. The graph proximity is described as the distance between two graphs. DynGPE can cluster similar climatic events in the embedding space. Abnormal climatic events are distant from most of the other events and can be detected using outlier detection methods. We conducted experiments by applying three outlier detection methods (i.e., isolation forest, local outlier factor, and box plot) to real meteorological data. The results showed that DynGPE achieves better results than the baseline by 44.3% on average in terms of the F-measure. Isolation forest provides the best performance and stability. It achieved higher results than the local outlier factor and box plot methods, namely, by 15.4% and 78.9% on average, respectively.

show abstract

DNETC: dynamic network embedding preserving both triadic closure evolution and community structures

Yang

Chen

et al. 2022

Knowl Inf Syst

View full text Add to dashboard Cite

Community-aware dynamic network embedding by using deep autoencoder

Cited by 13 publications

References 21 publications

Towards Robust Dynamic Network Embedding

Towards Robust Dynamic Network Embedding

Dynamic graph embedding for outlier detection on multiple meteorological time series

DNETC: dynamic network embedding preserving both triadic closure evolution and community structures

Contact Info

Product

Resources

About