Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking

Time series anomaly detection is a task that determines whether an unseen signal is normal or abnormal, and it is a crucial function in various real-world applications. Typical approach is to learn normal data representation using generative models, like Generative Adversarial Network (GAN), to disc...

Full description

Saved in:
Bibliographic Details
Main Authors: Ah-Hyung Shin, Seong Tae Kim, Gyeong-Moon Park
Format: Article
Language:English
Published: IEEE 2023-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10164104/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849233919278317568
author Ah-Hyung Shin
Seong Tae Kim
Gyeong-Moon Park
author_facet Ah-Hyung Shin
Seong Tae Kim
Gyeong-Moon Park
author_sort Ah-Hyung Shin
collection DOAJ
description Time series anomaly detection is a task that determines whether an unseen signal is normal or abnormal, and it is a crucial function in various real-world applications. Typical approach is to learn normal data representation using generative models, like Generative Adversarial Network (GAN), to discriminate between normal and abnormal signals. Recently, a few studies actively adopt Transformer to model time series data, but there is no pure Transformer-based GAN framework for time series anomaly detection. As a pioneer work, we propose a new pure Transformer-based GAN framework, called AnoFormer, and its effective training strategy for better representation learning. Specifically, we improve the detection ability of our model by introducing two-step masking strategies. The first step is Random masking: we design a random mask pool to hide parts of the signal randomly. This allows our model to learn the representation of normal data. The second step is Exclusive and Entropy-based Re-masking: we propose a novel refinement step to provide feedback to accurately model the exclusive and uncertain parts in the first step. We empirically demonstrate the effectiveness of re-masking step that generates more normal-like signals robustly. Extensive experiments on various datasets show that AnoFormer significantly outperforms the state-of-the-art methods in time series anomaly detection.
format Article
id doaj-art-a4acb3878c2e49d9b998250956f09001
institution Kabale University
issn 2169-3536
language English
publishDate 2023-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-a4acb3878c2e49d9b998250956f090012025-08-20T04:03:21ZengIEEEIEEE Access2169-35362023-01-0111740357404710.1109/ACCESS.2023.328992110164104Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step MaskingAh-Hyung Shin0https://orcid.org/0000-0003-1015-6167Seong Tae Kim1https://orcid.org/0000-0002-2132-6021Gyeong-Moon Park2https://orcid.org/0000-0003-4011-9981Department of Computer Science and Engineering, Kyung Hee University, Yongin-si, South KoreaDepartment of Computer Science and Engineering, Kyung Hee University, Yongin-si, South KoreaDepartment of Computer Science and Engineering, Kyung Hee University, Yongin-si, South KoreaTime series anomaly detection is a task that determines whether an unseen signal is normal or abnormal, and it is a crucial function in various real-world applications. Typical approach is to learn normal data representation using generative models, like Generative Adversarial Network (GAN), to discriminate between normal and abnormal signals. Recently, a few studies actively adopt Transformer to model time series data, but there is no pure Transformer-based GAN framework for time series anomaly detection. As a pioneer work, we propose a new pure Transformer-based GAN framework, called AnoFormer, and its effective training strategy for better representation learning. Specifically, we improve the detection ability of our model by introducing two-step masking strategies. The first step is Random masking: we design a random mask pool to hide parts of the signal randomly. This allows our model to learn the representation of normal data. The second step is Exclusive and Entropy-based Re-masking: we propose a novel refinement step to provide feedback to accurately model the exclusive and uncertain parts in the first step. We empirically demonstrate the effectiveness of re-masking step that generates more normal-like signals robustly. Extensive experiments on various datasets show that AnoFormer significantly outperforms the state-of-the-art methods in time series anomaly detection.https://ieeexplore.ieee.org/document/10164104/Anomaly detectionmaskingself-attentionsignal reconstructiontransformertime series analysis
spellingShingle Ah-Hyung Shin
Seong Tae Kim
Gyeong-Moon Park
Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
IEEE Access
Anomaly detection
masking
self-attention
signal reconstruction
transformer
time series analysis
title Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
title_full Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
title_fullStr Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
title_full_unstemmed Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
title_short Time Series Anomaly Detection Using Transformer-Based GAN With Two-Step Masking
title_sort time series anomaly detection using transformer based gan with two step masking
topic Anomaly detection
masking
self-attention
signal reconstruction
transformer
time series analysis
url https://ieeexplore.ieee.org/document/10164104/
work_keys_str_mv AT ahhyungshin timeseriesanomalydetectionusingtransformerbasedganwithtwostepmasking
AT seongtaekim timeseriesanomalydetectionusingtransformerbasedganwithtwostepmasking
AT gyeongmoonpark timeseriesanomalydetectionusingtransformerbasedganwithtwostepmasking