TadGAN: Time Series Anomaly Detection Using Generative Adversarial Networks
TadGAN: обнаружение аномалий во временных рядах с использованием генеративных состязательных сетей
2020-12-10
SCID: 54.1/8887qbw5
Discuss with AI
LSTM recurrent neural networksanomaly scoringcycle consistency lossgenerative adversarial networkstime series anomaly detection
Figures from the paper
Abstract (AI)
Time series anomalies can offer information relevant to critical situations facing various fields, from finance and aerospace to the IT, security, and medical domains. However, detecting anomalies in time series data is particularly challenging due to the vague definition o f a nomalies and said data's frequent lack of labels and highly complex temporal correlations. Current state-of-the-art unsupervised machine learning methods for anomaly detection suffer from scalability and portability issues, and may have high false positive rates. In this paper, we propose TadGAN, an unsupervised anomaly detection approach built on Generative Adversarial Networks (GANs). To capture the temporal correlations of time series distributions, we use LSTM Recurrent Neural Networks as base models for Generators and Critics. TadGAN is trained with cycle consistency loss to allow for effective time-series data reconstruction. We further propose several novel methods to compute reconstruction errors, as well as different approaches to combine reconstruction errors and Critic outputs to compute anomaly scores. To demonstrate the performance and generalizability of our approach, we test several anomaly scoring techniques and report the best-suited one. We compare our approach to 8 baseline anomaly detection methods on 11 datasets from multiple reputable sources such as NASA, Yahoo, Numenta, Amazon, and Twitter. The results show that our approach can effectively detect anomalies and outperform baseline methods in most cases (6 out of 11). Notably, our method has the highest averaged F1 score across all the datasets. Our code is open source and is available as a benchmarking tool.
Key Findings
1
Across 11 datasets from NASA, Yahoo, Numenta, Amazon, and Twitter, TadGAN outperformed baseline methods in 6 datasets and achieved the highest average F1 score.
2
Cycle-consistency loss enables TadGAN to reconstruct time-series data effectively for anomaly detection.
3
TadGAN is an unsupervised time-series anomaly detection method using LSTM-based GAN generators and critics to model complex temporal correlations.
4
The authors release TadGAN code as an open-source benchmarking tool.
5
The method introduces multiple reconstruction-error calculations and combines reconstruction errors with critic outputs to produce anomaly scores.
Research Object
multivariate and univariate time series data
Research Subject
unsupervised anomaly detection, including time-series reconstruction and anomaly-score computation
Publication Details
Publication Date
2020-12-10
Journal
Publisher
ISSN
Cited by
416
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai5
Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks2017
Anomaly detection: A survey2009
A Comparative Evaluation of Unsupervised Anomaly Detection Algorithms for Multivariate Data2016
Unsupervised real-time anomaly detection for streaming data2017
Time-Series Anomaly Detection Service at Microsoft2019