MAD-GAN: Multivariate Anomaly Detection for Time Series Data with Generative Adversarial Networks

Li, Dan; Chen, Dacheng; Jin, Baihong; Shi, Lei; Goh, Jonathan; Ng, See-Kiong

doi:10.48550/arxiv.1901.04997

Cited by 32 publications

(16 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first category of approaches analyzes each individual time-series by applying univariate models [1], [6]- [8], while the second one models multiple time-series as a unified entity [2], [4], [9]- [13]. From another perspective, existing anomaly detection models can also be categorized into two paradigms, namely forecasting-based models [2], [14], [15] and reconstructionbased models [4], [11]- [13]. In this section, we summarize important works about time-series anomaly detection and discuss these two paradigms in detail.…”

Section: Related Workmentioning

confidence: 99%

“…Generative Adversarial Networks (GANs) have also been widely used in multivariate time-series anomaly detection. Instead of treating each time-series independently, MAD-GAN [11] considers the entire variable set concurrently to capture the latent interactions among variables. GAN-Li [10] proposes a novel GAN-based anomaly detection method which deploys the GAN-trained discriminator together with the residuals between generator-reconstructed data and the actual samples.…”

Section: B Multivariate Anomaly Detectionmentioning

confidence: 99%

“…We compare MTAD-GAT with state-of-the-art models for multivariate time-series anomaly detection, including Omni-Anomaly [4], LSTM-NDT [2], KitNet [13], DAGMM [9], GAN-Li [10], MAD-GAN [11] and LSTM-VAE [12]. We use the same sliding window size n = 100 for all models.…”

Section: B Setupmentioning

confidence: 99%

See 2 more Smart Citations

Multivariate Time-series Anomaly Detection via Graph Attention Network

Zhao

Wang

Duan

et al. 2020

Preprint

View full text Add to dashboard Cite

Anomaly detection on multivariate time-series is of great importance in both data mining research and industrial applications. Recent approaches have achieved significant progress in this topic, but there is remaining limitations. One major limitation is that they do not capture the relationships between different time-series explicitly, resulting in inevitable false alarms. In this paper, we propose a novel self-supervised framework for multivariate time-series anomaly detection to address this issue. Our framework considers each univariate time-series as an individual feature and includes two graph attention layers in parallel to learn the complex dependencies of multivariate time-series in both temporal and feature dimensions. In addition, our approach jointly optimizes a forecasting-based model and a reconstruction-based model, obtaining better time-series representations through a combination of single-timestamp prediction and reconstruction of the entire time-series. We demonstrate the efficacy of our model through extensive experiments. The proposed method outperforms other state-of-the-art models on three real-world datasets. Further analysis shows that our method has good interpretability and is useful for anomaly diagnosis.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: B Multivariate Anomaly Detectionmentioning

confidence: 99%

See 1 more Smart Citation

Multivariate Time-series Anomaly Detection via Graph Attention Network

Zhao

Wang

Duan

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The ideas in [6], [13] are similar to that of in [7], [8], [14], [15] but uses a multivariate setting. The multivariate IDS solves the problem of deploying separate models for each sensor, thus can be deployed in a realtime resource-constrained environment.…”

Section: Background and Related Workmentioning

confidence: 99%

“…In attack scenario 2, the attacker's intent is to increase the value of the sensor abruptly by gradually increasing to a point well above the desired level. Figure 9 shows the reading of the sensor XMeas (6) with time when the sensor is under attack. Figure 8 and Figure 10 show the anomaly score at each time step, with the blue horizontal line as the threshold.…”

Section: Tennessee Eastman Process Datasetmentioning

confidence: 99%

Unsupervised Learning Based Robust Multivariate Intrusion Detection System for Cyber-Physical Systems using Low Rank Matrix

Dutta¹,

Mukhoty²,

Shukla³

2020

Preprint

View full text Add to dashboard Cite

Regular and uninterrupted operation of critical infrastructures such as power, transport, communication etc. are essential for proper functioning of a country. Cyber-attacks causing disruption in critical infrastructure service in the past, are considered as a significant threat. With the advancement in technology and the progress of the critical infrastructures towards IP based communication, cyber-physical systems are lucrative targets of the attackers. In this paper, we propose a robust multivariate intrusion detection system called RAD for detecting attacks in the cyber-physical systems in O(d) space and time complexity, where d is the number parameters in the system state vector. The proposed Intrusion Detection System(IDS) is developed in an unsupervised learning setting without using labelled data denoting attacks. It allows a fraction of the training data to be corrupted by outliers or under attack, by subscribing to robust training procedure. The proposed IDS outperforms existing anomaly detection techniques in several real-world datasets and attack scenarios.

show abstract