A Deep Neural Network for Unsupervised Anomaly Detection and Diagnosis in Multivariate Time Series Data

By Chuxu Zhang, Dongjin Song, Yuncong Chen, Xinyang Feng, Cristian Lumezanu, Wei Cheng, Jingchao Ni, Bo Zong, Haifeng Chen, Nitesh V. Chawla. AAAI Conference on Artificial Intelligence, 2018

934 citations recorded by the source corpus.

Nowadays, multivariate time series data are increasingly collected in various real world systems, e.g., power plants, wearable devices, etc. Anomaly detection and diagnosis in multivariate time series refer to identifying abnormal status in certain time steps and pinpointing the root causes. Building such a system, however, is challenging since it not only requires to capture the temporal dependency in each time series, but also need encode the inter-correlations between different pairs of time series. In addition, the system should be robust to noise and provide operators with different levels of anomaly scores based upon the severity of different incidents. Despite the fact that a number of unsupervised anomaly detection algorithms have been developed, few of them can jointly address these challenges. In this paper, we propose a Multi-Scale Convolutional Recurrent Encoder-Decoder (MSCRED), to perform anomaly detection and diagnosis in multivariate time series data. Specifically, MSCRED first constructs multi-scale (resolution) signature matrices to characterize multiple levels of the system statuses in different time steps. Subsequently, given the signature matrices, a convolutional encoder is employed to encode the inter-sensor (time series) correlations and an attention based Convolutional Long-Short Term Memory (ConvLSTM) network is developed to capture the temporal patterns. Finally, based upon the feature maps which encode the inter-sensor correlations and temporal information, a convolutional decoder is used to reconstruct the input signature matrices and the residual signature matrices are further utilized to detect and diagnose anomalies. Extensive empirical studies based on a synthetic dataset and a real power plant dataset demonstrate that MSCRED can outperform state-ofthe-art baseline methods.

Subfields: Artificial Intelligence, Computer Networks and Communications, Signal Processing

View Paper: https://www.semanticscholar.org/paper/d9cb449af41b6da76c98475ff268b3ef06d6dec1

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Signal Processing

Citations in the corpus

USAD: UnSupervised Anomaly Detection on Multivariate Time Series (2020, methodology): “Figure 2(B) presents the obtained results for five different window sizes K ∈ [5, 10, 20, 50, 100].”
A Transformer-based Framework for Multivariate Time Series Representation Learning (2020, methodology): “ries by directing learned representations to approximate a distance such as Dynamic Time Warping (DTW) between time series through a matrix factorization algorithm. A distinct approach is followed by Zhang et al. (2019), who use a composite convolutional - LSTM network with attention and a loss whic”
TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data (2022, methodology): “The OmniAnomaly, CAE-M and MSCRED models use sequential observations as input, allowing these methods to retain the temporal information.”
Deep Learning for Anomaly Detection (2020, methodology): “tional use of AEs with the exception that a particular network architecture tailored for a specific type of data is required to learn effective low-dimensional feature representations, such as CNN-AE [57, 173], LSTM-AE [98], Conv-LSTM-AE [94] and GCN (graph convolutional network)-AE [38]. This type ”
Deep Learning for Anomaly Detection in Time-Series Data: Review, Analysis, and Guidelines (2021, methodology): “However, there are unique properties of the time series structure, and there are several aspects of the environment in which the data are generated and analyzed that could affect the success of an anomaly detection algorithm.”
Anomaly Detection in Time Series: A Comprehensive Evaluation (2022, methodology): “Representatives of this category in our evaluation are: AutoEncoder (AE) [117], Bagel [79], DenoisingAutoEncoder (DAE) [117], Donut [150], EncDec-AD [88], FFT [111], Image-embedding CAE (IE-CAE) [44], LSTM-VAE [106], MSCRED [159], OmniAnomaly [125], PCI [157], PCC [121], RobustPCA [101], Spectral Re”
Timeseries Anomaly Detection using Temporal Hierarchical One-Class Network (2020, background): “Following [17, 28, 13, 31], we consider the unsupervised learning setting, and do not use any label information during training.”
Deep Learning for Time Series Anomaly Detection: A Survey (2022, neutral): “1) AE/DAE [151] T Semi P Point DAGMM [198] S Un P Point ✓ MSCRED [187] ST Un W ✓ Subseq USAD [13] T Un W Point APAE [72] T Un W RANSynCoders [4] ST Un P ✓ Point ✓ CAE-Ensemble [25] T Un W Subseq AMSL [192] T Self W -”
An Evaluation of Anomaly Detection and Diagnosis in Multivariate Time Series (2021, methodology): “The AE is the basis of many deep-learning based models for MVTS anomaly detection [7]–[9], thus we include various architectures for this.”
Potential, challenges and future directions for deep learning in prognostics and health management applications (2020, methodology): “Deep learning is an overarching concept that encompasses new variants of a range of established learning models, known as neural networks (Bishop, 2007), now more commonly referred to as deep neural networks (DNNs) (Goodfellow et al., 2016; Zhang et al., 2019a).”
Unsupervised Deep Anomaly Detection for Multi-Sensor Time-Series Signals (2021, methodology): “(9) Multi-scale convolutional recurrent encoder-decoder (MSCRED) [15], which is a reconstruction-based anomaly detection and diagnosis method.”