Applications of Anomaly Detection Using Deep Learning on Time Series Data

Nguyen, Van Du; Van, Linh; Kim, Jin‐Young; Kim, Kwangki; Kim, Jinsul

doi:10.1109/dasc/picom/datacom/cyberscitec.2018.00078

Cited by 12 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A graph neural network (GNN) considers correlations among sensors and hidden relationships within multivariate time series, and it has also found application in the domain of anomaly detection [14][15][16]. Besides these reconstruction-error-based methods, various neural network models have also been applied to build prediction-error-based detection methods, such as RNNs [17][18][19][20], CNNs [21], and GNNs [22,23]. Most existing anomaly detection methods are typically end-to-end models that predominantly output deterministic anomaly labels: classifying instances as either anomalies or not.…”

Section: Related Workmentioning

confidence: 99%

“…They lack the capability to indicate the confidence level or the probabilistic nature of their assessments, which is crucial in scenarios for which the distinction between normal and anomalous is not clear-cut. Moreover, the majority of existing research [8,9,13,18] focuses on enhancing the accuracy of detection while often neglecting the importance of providing explanations for the detected results. The authors of [24] introduced a novel design process that integrates human experts into the learning loop by presenting model outcomes in an understandable form.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

PrOuD: Probabilistic Outlier Detection Solution for Time-Series Analysis of Real-World Photovoltaic Inverters

He,

Huang,

Vogt

et al. 2023

Energies

View full text Add to dashboard Cite

Anomaly detection methods applied to time series are mostly viewed as black boxes that solely provide a deterministic answer for the detected target. Without a convincing explanation, domain experts can hardly trust the detection results and must conduct further time-series diagnoses in real-world applications. To overcome this challenge, we mathematically analyzed the sources of anomalies and novelties in multivariate time series as well as their relationships from the perspective of Gaussian-distributed non-stationary noise. Furthermore, we proposed mathematical methods to generate artificial time series and synthetic anomalies, with the goal of solving the problem of it being difficult to train and evaluate models for real-world applications due to the lack of sufficient data. In addition, we designed Probabilistic Outlier Detection (PrOuD), which is a general solution to provide interpretable detection results to assist domain experts with time-series analysis. PrOuD can convert the predictive uncertainty of a time-series value from a trained model into the estimated uncertainty of the detected outlier through Monte Carlo estimation. The experimental results obtained on both artificial time series and real-world photovoltaic inverter data demonstrated that the proposed solution can detect emerging anomalies accurately and quickly. The implemented PrOuD demo case shows its potential to make the detection results of existing detection methods more convincing so that domain experts can more efficiently complete their tasks, such as time-series diagnosis and anomalous pattern clustering.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

PrOuD: Probabilistic Outlier Detection Solution for Time-Series Analysis of Real-World Photovoltaic Inverters

He,

Huang,

Vogt

et al. 2023

Energies

View full text Add to dashboard Cite

show abstract

“…Therefore, the LSTM method is applied to identify the intrusions in a network system [59]. Various steps are carried out in the proposed method de􀅫ined in the table above [63]. As a result, the proposed LSTM model method is utilized for indicating results as the output classi􀅫ication, known as Binary Classi􀅫ication, which is used to detect normal and malicious activities, and the other is known as Multiclass Clas-si􀅫ication, which is used to give 􀅫ive outputs including R2L, Probe, Normal, DOS, U2R [54].…”

Section: B Working Of Lstmmentioning

confidence: 99%

“…, the results after the application of the confusion matrix in association with LSTM are shown by the classi􀅫ication of a binary and multiclass classi􀅫ier. Therefore, the paper results as giving 99.2% accuracy for the Binary classi􀅫ication model and 96.9% accuracy for the Multiclass classi􀅫ication model based on classi􀅫ications done by LSTM[54,63].…”

mentioning

confidence: 99%

A review on the impact of AI as big data in computer network environment

Ayub¹,

Mumtaz²,

Ahmad³

et al. 2021

J. adv. tec. eng. res.

View full text Add to dashboard Cite

The rapid growth of Arti􀅫icial Intelligence (AI) with the evolution of network technology has improved and progressed to meet the needs of people. Big data place an important role in the 􀅫ield of AI due to enhanced utilization in computer networks. Many applications are used for big data-based AI in a computer network environment. The problems and the attacks (such as Ransomware, Phishing attacks, and AI attacks) occurring in Network Technology are widely associated with AI as big data along with the solutions such as detection and identi􀅫ication of given attacks and malicious activities. The computer network technology environment is used on a large scale and varies widely with the use of AI in the context of Big Data. However, more advanced technologies are applied with AI to ensure the security of Computer networks. Due to the vast amount of traf􀅫ic occurring in networks, the result leads to the occurrence of damage in the network, and intrusion detection is a process that proved to be the most essential to ensure the security provided to the system. The de􀅫ined attacks are speci􀅫ied as the main issues and will prevent vulnerable and malicious activities. The network requires a system to detect intrusions, which is further used to observe and identify the system and networks for anomalies and malicious intrusions as Intrusion Detection System (IDS), Firewall, Encryption, and Anti-spam technologies are used to identify any malicious activity and prevent it. For these attacks, Long Short Term Memory (LSTM) is used in deep learning, and many other technologies such as Machine learning and Data mining are associated with LSTM to detect intrusions. The computer networks with the application of big data as AI were made highly secure and ef􀅫icient systems by implementing the proposed LSTM method.

show abstract

“…In order to overcome the drawback of hand-crafted features, many models based on deep appearance features are proposed to detect abnormal events. The above deep appearance features can be obtained by using convolutional neural networks [36], [37], recurrent neural networks [38], [39] and autoencoder networks [40], [41].…”

Section: B Deep Appearance Features-based Modelsmentioning

confidence: 99%

Abnormal Event Detection via Feature Expectation Subgraph Calibrating Classification in Video Surveillance Scenes

Deng

et al. 2020

IEEE Access

View full text Add to dashboard Cite

At present, the existing abnormal event detection models based on deep learning mainly focus on data represented by a vectorial form, which pay little attention to the impact of the internal structure characteristics of feature vector. In addition, a single classifier is difficult to ensure the accuracy of classification. In order to address the above issues, we propose an abnormal event detection hybrid modulation method via feature expectation subgraph calibrating classification in video surveillance scenes in this paper. Our main contribution is to calibrate the classification of a single classifier by constructing feature expectation subgraphs. First, we employ convolutional neural network and long short-term memory models to extract the spatiotemporal features of video frame, and then construct the feature expectation subgraph for each key frame of every video, which could be used to capture the internal sequential and topological relational characteristics of structured feature vector. Second, we project expectation subgraphs on the sparse vector to combine with a support vector classifier to calibrate the results of a linear support vector classifier. Finally, the experiments on a common dataset named UCSDped1 and a coal mining video dataset in comparison with some existing works demonstrate that the performance of the proposed method is better than several the state-of-the-art approaches. INDEX TERMS Abnormal event detection, feature expectation subgraph, calibrating classification, sequential and topological relational characteristics.

show abstract

Applications of Anomaly Detection Using Deep Learning on Time Series Data

Cited by 12 publications

References 6 publications

PrOuD: Probabilistic Outlier Detection Solution for Time-Series Analysis of Real-World Photovoltaic Inverters

PrOuD: Probabilistic Outlier Detection Solution for Time-Series Analysis of Real-World Photovoltaic Inverters

A review on the impact of AI as big data in computer network environment

Abnormal Event Detection via Feature Expectation Subgraph Calibrating Classification in Video Surveillance Scenes

Contact Info

Product

Resources

About