Companion Proceedings of the Web Conference 2020 2020
DOI: 10.1145/3366424.3391266
|View full text |Cite
|
Sign up to set email alerts
|

ASA: Adversary Situation Awareness via Heterogeneous Graph Convolutional Networks

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
33
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
5
1
1

Relationship

0
7

Authors

Journals

citations
Cited by 35 publications
(33 citation statements)
references
References 7 publications
0
33
0
Order By: Relevance
“…GeniePath [23] learns convolutional layers and neighbor weights using LSTM and the attention mechanism [34]. GEM [24], SemiGNN [37], ASA [41], and Player2Vec [48] all construct multiple homo-graphs based on node relations in corresponding datasets. After aggregating neighborhood information with GNNs on each homo-graph, SemiGNN and Player2Vec adopt attention mechanism to aggregate node embeddings across multiple homo-graphs; while GEM learns weighting parameters for different homo-graphs, and ASA directly sums information from each homo-graph.…”
Section: Related Workmentioning
confidence: 99%
“…GeniePath [23] learns convolutional layers and neighbor weights using LSTM and the attention mechanism [34]. GEM [24], SemiGNN [37], ASA [41], and Player2Vec [48] all construct multiple homo-graphs based on node relations in corresponding datasets. After aggregating neighborhood information with GNNs on each homo-graph, SemiGNN and Player2Vec adopt attention mechanism to aggregate node embeddings across multiple homo-graphs; while GEM learns weighting parameters for different homo-graphs, and ASA directly sums information from each homo-graph.…”
Section: Related Workmentioning
confidence: 99%
“…Among the above works, few works [4]- [6] have noticed the camouflage behaviors of fraudsters. All these works can only handle a multi-relation graph, where all nodes are considered to be the same type.…”
Section: B Graph Based Fraud Detectionmentioning
confidence: 99%
“…All these works can only handle a multi-relation graph, where all nodes are considered to be the same type. ASA [6] creates static features for directly aggregating messages in each homogeneous graph. GraphConsis [4] suffers from inflexible filtering thresholds and unsupervised similarity measures.…”
Section: B Graph Based Fraud Detectionmentioning
confidence: 99%
See 2 more Smart Citations