Is a Single Embedding Enough? Learning Node Representations that Capture Multiple Social Contexts

Epasto, Alessandro; Perozzi, Bryan

doi:10.1145/3308558.3313660

Cited by 100 publications

(117 citation statements)

References 42 publications

Supporting

Mentioning

114

Contrasting

Unclassified

Order By: Relevance

“…The recently emerged research field of graph embedding [16][17][18][19][20] enables low-dimensional representations of large-scale networks, thereby ensuring tractable visualization and analysis of the resulting networks. Domains like sociology and biology have successfully applied these techniques to perform a variety of tasks such as node classification, link prediction and clustering [21][22][23] .…”

Section: Reducing the Complexity Of Financial Network Using Network mentioning

confidence: 99%

Reducing the complexity of financial networks using network embeddings

Boersma

Maliutin

Sourabh

et al. 2020

Sci Rep

View full text Add to dashboard Cite

Accounting scandals like Enron (2001) and Petrobas (2014) remind us that untrustworthy financial information has an adverse effect on the stability of the economy and can ultimately be a source of systemic risk. This financial information is derived from processes and their related monetary flows within a business. But as the flows are becoming larger and more complex, it becomes increasingly difficult to distill the primary processes for large amounts of transaction data. However, by extracting the primary processes we will be able to detect possible inconsistencies in the information efficiently. We use recent advances in network embedding techniques that have demonstrated promising results regarding node classification problems in domains like biology and sociology. We learned a useful continuous vector representation of the nodes in the network which can be used for the clustering task, such that the clusters represent the meaningful primary processes. The results show that we can extract the relevant primary processes which are similar to the created clusters by a financial expert. Moreover, we construct better predictive models using the flows from the extracted primary processes which can be used to detect inconsistencies. Our work will pave the way towards a more modern technology and data-driven financial audit discipline.

show abstract

Section: Reducing the Complexity Of Financial Network Using Network mentioning

confidence: 99%

Reducing the complexity of financial networks using network embeddings

Boersma

Maliutin

Sourabh

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…The third type is the bipartite graph, and there have been many outstanding studies of bipartite graph embedding [28], [36], [37]. Graph embedding has a wide range of applications, such as classification [13], [14], [33], clustering [37], [38], recommendation [28], [36], link prediction [39]- [41], visualization [28], [42], anomaly detection [35] and face recognition [43].…”

Section: B Graph Embeddingmentioning

confidence: 99%

DBGE: Employee Turnover Prediction Based on Dynamic Bipartite Graph Embedding

et al. 2020

View full text Add to dashboard Cite

The issue of employee turnover is always critical for companies, and accurate predictions can help them prepare in time. Most past studies on employee turnover have focused on analyzing impact factors or using simple network centrality measures. In this paper, we study the problem from a completely new perspective by modeling users' historical job records as a dynamic bipartite graph. Specifically, we propose a bipartite graph embedding method with temporal information called dynamic bipartite graph embedding (DBGE) to learn the vector representation of employees and companies. Our approach not only considers the relations between employees and companies but also incorporates temporal information embedded in consecutive work records. We first define the Horary Random Walk on a bipartite graph to generate a sequence for each vertex in chronological order. Then, we employ the skip-gram model to obtain a temporal low-dimensional vector representation for each vertex and apply machine learning methods to predict employee turnover behavior by combining embedded features with employees' basic information. Experiments on a real-world dataset collected from one of China's largest online professional social networks show that the features learned through DBGE can significantly improve turnover prediction performance. Moreover, experiments on public Amazon and Taobao datasets show that our approach achieves better performance in the link prediction and visualization task than other graph embedding methods that do not consider temporal information. INDEX TERMS Graph embedding, employee turnover prediction, dynamic bipartite graph, horary random walk.

show abstract

“…Besides, GraphSAGE [38] allows embedding vectors to be efficiently generated for unseen nodes instead of training embedding vectors of all nodes. There are also other algorithms, such as Splitter [39] which learns multiple embeddings.…”

Section: ) Graph Embeddingmentioning

confidence: 99%

Complex Network Metrics: Can Deep Learning Keep up With Tailor-Made Reference Algorithms?

Wandelt

Shi²,

Sun

2020

IEEE Access

View full text Add to dashboard Cite

Complex network metrics are used for ranking the importance of nodes in many different applications, e.g., spreader identification and percolation analysis. Examples for such node metrics include the degree centrality and betweenness centrality. The computation of some of these metrics is computationally expensive, e.g. the computation of betweenness takes O(N 3) of computation steps in the worst case, where N is the number of nodes. In this study, we investigate the ability of deep learning via graph embedding to predict node centrality metrics in complex networks. Our study reveals that deep learning can identify vital nodes well for several of these metrics. Compared to exact computation with prohibitive costs, deep learning can get significant speedups, which scale up well with the size of the network. Further investigation on betweenness centrality reveals that the accuracy of deep learning is not as good as some tailor-made, tuned approximation algorithms. However, deep learning offers a nice trade-off between runtime and quality with its linear computational complexity regarding the network size, which makes it scalable on large networks and especially for these metrics with expensive computational costs. Our work is the first to explore the ability of deep learning on a wide range of networks metrics, and will hopefully induce future work on improved graph embeddings, tuned for specific network metrics. INDEX TERMS Complex networks, network metrics, deep learning.

show abstract

Is a Single Embedding Enough? Learning Node Representations that Capture Multiple Social Contexts

Cited by 100 publications

References 42 publications

Reducing the complexity of financial networks using network embeddings

Reducing the complexity of financial networks using network embeddings

DBGE: Employee Turnover Prediction Based on Dynamic Bipartite Graph Embedding

Complex Network Metrics: Can Deep Learning Keep up With Tailor-Made Reference Algorithms?

Contact Info

Product

Resources

About