BeatGAN: Anomalous Rhythm Detection using Adversarially Generated Time Series

By Bin Zhou, Shenghua Liu, Bryan Hooi, Xueqi Cheng, Jing Ye. International Joint Conference on Artificial Intelligence, 2019

368 citations recorded by the source corpus.

Given a large-scale rhythmic time series containing mostly normal data segments (or `beats'), can we learn how to detect anomalous beats in an effective yet efficient way? For example, how can we detect anomalous beats from electrocardiogram (ECG) readings? Existing approaches either require excessively high amounts of labeled and balanced data for classification, or rely on less regularized reconstructions, resulting in lower accuracy in anomaly detection. Therefore, we propose BeatGAN, an unsupervised anomaly detection algorithm for time series data. BeatGAN outputs explainable results to pinpoint the anomalous time ticks of an input beat, by comparing them to adversarially generated beats. Its robustness is guaranteed by its regularization of reconstruction error using an adversarial generation approach, as well as data augmentation using time series warping. Experiments show that BeatGAN accurately and efficiently detects anomalous beats in ECG time series, and routes doctors' attention to anomalous time ticks, achieving accuracy of nearly 0.95 AUC, and very fast inference (2.6 ms per beat). In addition, we show that BeatGAN accurately detects unusual motions from multivariate motion-capture time series data, illustrating its generality.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/715008d75d97a0a975bbd644a977f83942c5eb6a

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

1. Mingsheng Long
2. Jianmin Wang
3. Qingsong Wen
4. Haixu Wu
5. Sungroh Yoon
6. Jiehui Xu
7. Jihun Yi
8. Kukjin Choi

In Signal Processing

1. Mingsheng Long
2. Jianmin Wang
3. Liang Sun
4. Haixu Wu
5. Tian Zhou
6. Qingsong Wen
7. Xue Wang
8. Jiehui Xu

Citations in the corpus

• An Evaluation of Anomaly Detection and Diagnosis in Multivariate Time Series (2021, methodology): “While the good performance of UAE could be partly due to the prevalence of temporal anomalies, our study makes it clear that recently proposed deep algorithms [4], [5], [9], [10], [13] fail to effectively detect even these simple [49] anomalies in MVTS datasets.”
• Deep Learning for Time Series Anomaly Detection: A Survey (2022, neutral): “3) MAD-GAN [113] ST Un W Subseq BeatGAN [195] T Un W Subseq DAEMON [37] T Un W ✓ Subseq FGANomaly [56] T Un W Point + Subseq DCT-GAN* [116] T Un W -”
• Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancy (2021, methodology): “GANs (Goodfellow et al., 2014) are also used for reconstruction-based anomaly detection (Schlegl et al., 2019; Li et al., 2019a; Zhou et al., 2019) and perform as an adversarial regularization.”
• Timeseries Anomaly Detection using Temporal Hierarchical One-Class Network (2020, background): “Other more powerful deep generative models, such as the recurrent variational autoencoder [28], and variants of the generative adversarial network (GAN) (e.g., MAD-GAN [13] and BeatGAN [33]) have also been proposed.”
• Graph Neural Network-Based Anomaly Detection in Multivariate Time Series (2021, background): “…(CNN) based models (Munir et al. 2018), Long Short Term Memory (LSTM) (Filonov, Lavrentyev, and Vorontsov 2016; Hund-man et al. 2018a; Park, Hoshi, and Kemp 2018) and Generative Adversarial Networks (GAN) models (Zhou et al. 2019; Li et al. 2019), have found success in practical time series tasks.”
• Time Series Data Augmentation for Deep Learning: A Survey (2020, methodology): “However, it does not review the data augmentation methods for other common tasks like time series forecasting [Bandara et al. , 2020; Hu et al. , 2020; Lee and Kim, 2020] and anomaly detection [Lim et al. , 2018; Zhou et al. , 2019; Gao et al. , 2020].”
• 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.”
• DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detection (2023, methodology): “Recent works in time series anomaly detection also include generative adversarial networks (GANs) based methods [16, 16, 41, 87] and deep reinforcement learning (DRL) based methods [30, 80].”