2020
DOI: 10.1109/tsusc.2018.2793284
|View full text |Cite
|
Sign up to set email alerts
|

Deep Learning Based Multi-Channel Intelligent Attack Detection for Data Security

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
72
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
6
3
1

Relationship

3
7

Authors

Journals

citations
Cited by 190 publications
(82 citation statements)
references
References 40 publications
0
72
0
Order By: Relevance
“…The STL technique consists of two stages for the classification, including, 1) feature representation is learned from a large collection of unlabeled data and 2) apply the first stage to labeled data for the classification task. The recurrent neural network algorithms are used by IDS models, such as long short-term memory (LSTM) [44], [49], forward propagation and back propagation (FPBP) [45], and gated recurrent unit (GRU) [47]. Therefore, many existing IDS systems utilize three datasets, including, KDD Cup 1999 dataset, NSL-KDD dataset, and UNSW-NB15 dataset, which are outdated and of very limited practical value for a modern IDS.…”
Section: Evaluation Resultsmentioning
confidence: 99%
“…The STL technique consists of two stages for the classification, including, 1) feature representation is learned from a large collection of unlabeled data and 2) apply the first stage to labeled data for the classification task. The recurrent neural network algorithms are used by IDS models, such as long short-term memory (LSTM) [44], [49], forward propagation and back propagation (FPBP) [45], and gated recurrent unit (GRU) [47]. Therefore, many existing IDS systems utilize three datasets, including, KDD Cup 1999 dataset, NSL-KDD dataset, and UNSW-NB15 dataset, which are outdated and of very limited practical value for a modern IDS.…”
Section: Evaluation Resultsmentioning
confidence: 99%
“…In recent years, CNN based approaches retain the overwhelming benefits compared to some previous handcrafted features extraction methods, particularly in semantics representation of image [12,34,35]. In the deep learning based approaches, the features of the earlier layers contain higher spatial resolution for precise local features, while the features in later layers indicates more semantic or global information [36].…”
Section: B Proposed Facial Features Extractionmentioning
confidence: 99%
“…In this paper, we calculate the angle cosine value which reflects the similarity between documents and cluster documents with DBSCAN. DBSCAN, proposed by Martin Ester in 1996, is a density-based clustering algorithm which is widely cited in scientific literature [23], and it is awarded the test of time award in 2014 [24]. When clustering, other thanmeans, DBSCAN does not need to specify the number of clusters and it can find the clusters of arbitrary shape.…”
Section: Clustering Of Textual Documentsmentioning
confidence: 99%