Anomaly detection in smart manufacturing: An Adaptive Adversarial Transformer-based model

IF 12.2 1区 工程技术 Q1 ENGINEERING, INDUSTRIAL Journal of Manufacturing Systems Pub Date : 2024-10-22 DOI:10.1016/j.jmsy.2024.09.021
Moussab Orabi , Kim Phuc Tran , Philipp Egger , Sébastien Thomassey
{"title":"Anomaly detection in smart manufacturing: An Adaptive Adversarial Transformer-based model","authors":"Moussab Orabi ,&nbsp;Kim Phuc Tran ,&nbsp;Philipp Egger ,&nbsp;Sébastien Thomassey","doi":"10.1016/j.jmsy.2024.09.021","DOIUrl":null,"url":null,"abstract":"<div><div>In Industry 5.0, smart manufacturing brings additional intricacies and novel data processing challenges. Given the evolving nature of manufacturing processes and the inherent complexity of data, including noise and missing entries, achieving accurate anomaly detection becomes even more intricate. Conventional methods often miss nuanced anomalies, especially when dealing with high-dimensional, multivariate, non-stationary data. These data types are typical of smart manufacturing environments. Hence, many recent approaches have embraced deep learning to confront these challenges, making use of diverse attention mechanisms to acquire data representations. However, in manufacturing, where the dynamics of time series data change over time, methods relying solely on pointwise or pairwise representations often fall short. Thus, ensuring product quality and operational integrity calls for even more advanced methodologies. The deficiency lies in the capability of state-of-the-art models to effectively capture abnormal patterns while considering both local and global contextual information. This challenge is compounded by the rarity of anomalies, making it exceedingly challenging to establish substantial associations between individual abnormal points and the entire time series. To tackle these challenges, we introduce the “<strong>A</strong>daptive <strong>A</strong>dversarial <strong>T</strong>ransformer” as a novel deep learning technique that combines Transformer architecture with an anomaly attention mechanism and Adversarial Learning. Our Model effectively captures intricate temporal patterns, distinguishes normal and anomalous behaviors, and dynamically adjusts thresholds to align with the evolving dynamics of time-series data. Empirical validation on four benchmark datasets and three real-world manufacturing datasets demonstrates our model’s effectiveness compared to the state-of-the-art, as evidenced by the F1-Score.</div></div>","PeriodicalId":16227,"journal":{"name":"Journal of Manufacturing Systems","volume":"77 ","pages":"Pages 591-611"},"PeriodicalIF":12.2000,"publicationDate":"2024-10-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Manufacturing Systems","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0278612524002255","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, INDUSTRIAL","Score":null,"Total":0}
引用次数: 0

Abstract

In Industry 5.0, smart manufacturing brings additional intricacies and novel data processing challenges. Given the evolving nature of manufacturing processes and the inherent complexity of data, including noise and missing entries, achieving accurate anomaly detection becomes even more intricate. Conventional methods often miss nuanced anomalies, especially when dealing with high-dimensional, multivariate, non-stationary data. These data types are typical of smart manufacturing environments. Hence, many recent approaches have embraced deep learning to confront these challenges, making use of diverse attention mechanisms to acquire data representations. However, in manufacturing, where the dynamics of time series data change over time, methods relying solely on pointwise or pairwise representations often fall short. Thus, ensuring product quality and operational integrity calls for even more advanced methodologies. The deficiency lies in the capability of state-of-the-art models to effectively capture abnormal patterns while considering both local and global contextual information. This challenge is compounded by the rarity of anomalies, making it exceedingly challenging to establish substantial associations between individual abnormal points and the entire time series. To tackle these challenges, we introduce the “Adaptive Adversarial Transformer” as a novel deep learning technique that combines Transformer architecture with an anomaly attention mechanism and Adversarial Learning. Our Model effectively captures intricate temporal patterns, distinguishes normal and anomalous behaviors, and dynamically adjusts thresholds to align with the evolving dynamics of time-series data. Empirical validation on four benchmark datasets and three real-world manufacturing datasets demonstrates our model’s effectiveness compared to the state-of-the-art, as evidenced by the F1-Score.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
智能制造中的异常检测:基于自适应对抗变换器的模型
在工业 5.0 中,智能制造带来了更多的复杂性和新的数据处理挑战。鉴于制造流程不断变化的性质以及数据固有的复杂性(包括噪声和缺失条目),实现准确的异常检测变得更加复杂。传统方法往往会遗漏细微的异常情况,尤其是在处理高维、多变量、非稳态数据时。这些数据类型是智能制造环境的典型特征。因此,最近的许多方法都采用了深度学习来应对这些挑战,利用不同的注意机制来获取数据表征。然而,在制造业中,时间序列数据的动态会随着时间的推移而发生变化,仅仅依靠点或成对表示的方法往往无法满足要求。因此,确保产品质量和操作完整性需要更先进的方法。不足之处在于,最先进的模型无法在考虑局部和全局背景信息的同时有效捕捉异常模式。异常情况的罕见性加剧了这一挑战,使得在单个异常点和整个时间序列之间建立实质性关联变得极具挑战性。为了应对这些挑战,我们引入了 "自适应对抗变换器 "作为一种新型深度学习技术,它将变换器架构与异常关注机制和对抗学习相结合。我们的模型能有效捕捉错综复杂的时间模式,区分正常和异常行为,并动态调整阈值以适应时间序列数据不断变化的动态。在四个基准数据集和三个现实世界制造业数据集上的经验验证表明,我们的模型与最先进的模型相比非常有效,F1 分数就是证明。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Journal of Manufacturing Systems
Journal of Manufacturing Systems 工程技术-工程:工业
CiteScore
23.30
自引率
13.20%
发文量
216
审稿时长
25 days
期刊介绍: The Journal of Manufacturing Systems is dedicated to showcasing cutting-edge fundamental and applied research in manufacturing at the systems level. Encompassing products, equipment, people, information, control, and support functions, manufacturing systems play a pivotal role in the economical and competitive development, production, delivery, and total lifecycle of products, meeting market and societal needs. With a commitment to publishing archival scholarly literature, the journal strives to advance the state of the art in manufacturing systems and foster innovation in crafting efficient, robust, and sustainable manufacturing systems. The focus extends from equipment-level considerations to the broader scope of the extended enterprise. The Journal welcomes research addressing challenges across various scales, including nano, micro, and macro-scale manufacturing, and spanning diverse sectors such as aerospace, automotive, energy, and medical device manufacturing.
期刊最新文献
Material removal rate optimization with bayesian optimized differential evolution based on deep learning in robotic polishing Leveraging AI for energy-efficient manufacturing systems: Review and future prospectives Investigation of assistance systems in assembly in the context of digitalization: A systematic literature review Machining parameter optimization for a batch milling system using multi-task deep reinforcement learning A dynamic artificial bee colony for fuzzy distributed energy-efficient hybrid flow shop scheduling with batch processing machines
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1