检索结果-内蒙古大学图书馆

A Multi-scale Patch Mixer Network for time series anomaly detection

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2025年 140卷

作者： Wang, Qiushi Zhu, Yueming Sun, Zhicheng Li, Dong Ma, Yunbin Chinese Acad Sci Key Lab Networked Control Syst Shenyang 110016 Peoples R China Chinese Acad Sci Shenyang Inst Automat Shenyang 110016 Peoples R China PipeChina Inst Sci & Technol Langfang 065000 Peoples R China

With the development of Internet of Things (IoT) technology, a large amount of data with temporal characteristics is collected and stored. How to efficiently and accurately identify anomalies from these data is a major challenge. At present, there are many problems in the application of anomaly detection, including non- stationary data, complex and difficult-to-collect anomalies, the need for real-time detection and the limitation of computing resources. But few methods can comprehensively consider these issues. To overcome these challenges, we propose a lightweight neural network, Multi-scale Patch Mixer Network (MP-MixerNet). It is mainly composed of a Mixer Block based on fully connected layer design, which contains a Temporal-Mixer and a Spatial-Mixer, and can simultaneously model the intra- and inter-series dependencies of multivariate time series. We also perform multi-scale patch segmentation based on frequency analysis, which helps the model extract robust features from multiple period views. In addition, we design an Input Stabilization module to help the model deal with data distribution shift. Experimental results on a public time series anomaly detection dataset show that we are able to achieve higher comprehensive performance with fewer parameters and inference time.

关键词： Lightweight Multi-scale time series anomaly detection Fully connected layer

来源：评论

学校读者我要写书评

暂无评论

Decomposition-based multi-scale transformer framework for time series anomaly detection

引用

NEURAL NETWORKS 2025年 187卷 107399页

作者： Zhang, Wenxin Luo, Cuicui Univ Chinese Acad Sci Sch Comp Sci & Technol Beijing 100000 Peoples R China Univ Chinese Acad Sci Int Coll Beijing 100000 Peoples R China

time series anomaly detection is crucial for maintaining stable systems. Existing methods face two main challenges. First, it is difficult to directly model the dependencies of diverse and complex patterns within the sequences. Second, many methods that optimize parameters using mean squared error struggle with noise in the time series, leading to performance deterioration. To address these challenges, we propose a transformer based framework built on decomposition (TransDe) for multivariate time series anomaly detection. The key idea is to combine the strengths of time series decomposition and transformers to effectively learn the complex patterns in normal time series data. A multi-scale patch-based transformer architecture is proposed to exploit the representative dependencies of each decomposed component of the time series. Furthermore, a contrastive learn paradigm based on patch operation is proposed, which leverages KL divergence to align the positive pairs, namely the pure representations of normal patterns between different patch-level views. A novel asynchronous loss function with a stop-gradient strategy is further introduced to enhance the performance of TransDe effectively. It can avoid time-consuming and labor-intensive computation costs in the optimization process. Extensive experiments on five public datasets are conducted and TransDe shows superiority compared with twelve baselines in terms of F1 score. Our code is available at https://***/shaieesss/TransDe.

关键词： time series anomaly detection Unsupervised learning Neural networks

来源：评论

学校读者我要写书评

暂无评论

Generality-aware self-supervised transformer for multivariate time series anomaly detection

引用

APPLIED INTELLIGENCE 2025年第7期55卷 1-15页

作者： Cho, Yucheol Lee, Jae-Hyeok Ham, Gyeongdo Jang, Donggon Kim, Dae-shik Korea Adv Inst Sci & Technol Sch Elect Engn Daejeon 34141 South Korea

Efficient identification of anomalies within multivariate time series data holds significant relevance in contemporary industrial settings. The challenge lies in swiftly and accurately pinpointing anomalous data points. This challenge is further compounded by factors such as the absence of labeled anomalies, data volatility, and the need for ultra-fast inference times. While previous approaches have introduced advanced deep learning models to address these challenges, comprehensive efforts to tackle all these issues simultaneously have been limited. Recent developments in unsupervised learning-based models have demonstrated remarkable performance. However, many of these models rely on reconstruction error as an anomaly score, making them sensitive to unseen normal data patterns. To address this limitation, we propose a novel framework, generality-aware self-supervised transformer for multivariate time series anomaly detection, which utilizes a transformer that effectively generalizes normal data patterns through self-knowledge distillation. Furthermore, we incorporate an auxiliary decoder to compute generality-based anomaly scores, thereby enhancing the differentiation between anomalous and normal data points in testing datasets. In our study, encompassing a diverse range of publicly available datasets and our own extracted data from linear motion (LM) guides and reducers built to model the vertical and rotational motions of robots, we establish the superior anomaly detection performance of our framework compared to existing state-of-the-art models. Notably, we verify that this improved performance is achieved while also considering time efficiency.

关键词： time series anomaly detection Transformer Generalization Self-supervised learning Self-knowledge distillation

来源：评论

学校读者我要写书评

暂无评论

Uncertainty-informed dynamic threshold for time series anomaly detection

引用

EXPERT SYSTEMS WITH APPLICATIONS 2025年 278卷

作者： Lee, Jungmin Lee, Jiyoon Kim, Seoung Bum Korea Univ Dept Ind & Management Engn 145 Anam Ro Seoul 02841 South Korea SK Innovat Optimizat & Analyt Off 99 Jong Ro Seoul 03188 South Korea

As time series data continues to be collected across various fields, the importance of automated anomaly detection systems is steadily increasing. A key challenge in anomaly detection lies in setting an optimal threshold for anomaly scores to distinguish anomalies from normal data. Most existing studies use a fixed threshold, often resulting in misclassification of ambiguous data. Therefore, defining a dynamic and optimal threshold is crucial for improving detection performance. We aim to quantify uncertainty as a metric that determines the degree of ambiguity in the data. Because our models are trained only on normal data, anomalies exhibiting patterns divergent from the normal data entail higher uncertainty. Accordingly, in this study, we propose a dynamic thresholding method that better aligns with the nature of the data through uncertainty quantification. Through experimentation with synthetic datasets and five benchmark datasets for time series anomaly detection, we demonstrate the efficacy of our proposed method. Our proposed method outperforms both the fixed threshold and existing dynamic thresholding methods, achieving an average F1-score improvement of over 0.06 across benchmark datasets. In particular, the performance improvement is more significant when the distributions of normal data and anomalies are more similar. The source code can be accessed at https://***/jungminkr9195/UDT.

关键词： Dynamic threshold time series anomaly detection Uncertainty quantification

来源：评论

学校读者我要写书评

暂无评论

Retentive network-based time series anomaly detection in cyber-physical systems

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2025年 145卷

作者： Min, Zhaoyi Xiao, Qianqian Abbas, Muhammad Zhang, Duanjin Zhengzhou Univ Sch Elect & Informat Engn Zhengzhou 450001 Peoples R China Zhejiang Yuexiu Univ Eit Data Sci & Commun Coll Shaoxing 312000 Peoples R China Zhejiang Yuexiu Univ New Zealand Res Ctr Shaoxing 312000 Peoples R China

time series data are ubiquitous in the operation of cyber-physical systems (CPS), encompassing network traffic data, sensor measurements, and other relevant data streams. Intelligent anomaly detection methods are crucial for identifying deviations in these time series, which can enhance system maintainability and reliability. In this study, we propose an unsupervised reconstruction-based anomaly detection method utilizing retentive network (RetNet), a novel variant of Transformer. A one-dimensional convolutional neural network is employed to map the raw time series into the RetNet model space, analogous to word embedding techniques in natural language processing. The rotary position embedding introduced in RetNet can simultaneously incorporate absolute and relative positional information, which is beneficial for modeling temporal dependency in time series. Meanwhile, the multi-scale retention mechanism of RetNet facilitates the learning of informative representations of the dominant normal patterns from training data. The observed anomalies result in larger reconstruction errors, which are subsequently detected by the peaks-over-threshold (POT) method using a dynamic threshold. We evaluate the proposed method on four benchmark datasets. Experimental results demonstrate that the proposed method outperforms the best baseline by 4.07% in terms of the F1 score.

关键词： Retentive network time series anomaly detection Convolution neural network Transformer Dynamic threshold Cyber-physical systems

来源：评论

学校读者我要写书评

暂无评论

Deep time series anomaly detection with Local Temporal Pattern Learning

Deep Time Series Anomaly Detection with Local Temporal Patte...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Li, Yizhou Wang, Yijie Xu, Hongzuo Zhou, Xiaohui National Key Laboratory of Parallel and Distributed Computing College of Computer Science and Technology National University of Defense Technology Changsha410073 China Beijing100091 China

ISBN: (纸本)9798350368741

Self-supervised time series anomaly detection (TSAD) demonstrates remarkable performance improvement by extracting high-level data semantics through proxy tasks. Nonetheless, most existing self-supervised TSAD techniques rely on manual- or neural-based transformations when designing proxy tasks, overlooking the intrinsic temporal patterns of time series. This paper proposes a local temporal pattern learning-based time series anomaly detection (LTPAD). LTPAD first generates sub-sequences. Pairwise sub-sequences naturally manifest proximity relationships along the time axis, and such correlations can be used to construct supervision and train neural networks to facilitate the learning of temporal patterns. time intervals between two sub-sequences serve as labels for sub-sequence pairs. By classifying these labeled data pairs, our model captures the local temporal patterns of time series, thereby modeling the temporal pattern-aware "normality". Abnormal scores of testing data are acquired by evaluating their conformity to these learned patterns shared in training data. Extensive experiments show that LTPAD significantly outperforms state-of-the-art competitors. © 2025 IEEE.

关键词： Local Temporal Pattern Self-supervised Learning time series anomaly detection

来源：评论

学校读者我要写书评

暂无评论

MulGad: Multi-granularity contrastive learning for multivariate time series anomaly detection

引用

INFORMATION FUSION 2025年 119卷

作者： Xiao, Bo-Wen Xing, Hong-Jie Li, Chun-Guo Hebei Univ Coll Math & Informat Sci Hebei Key Lab Machine Learning & Computat Intellig Baoding 071002 Peoples R China Hebei Univ Coll Cyber Secur & Comp Baoding 071002 Peoples R China

Since the normal patterns of time series change dynamically over time, unsupervised time series anomaly detection methods have to face the overfitting problem. Although some approaches based on contrastive learning try to solve this problem, these methods ignore the complex intrinsic correlations within the given time series and may face the representation collapse problem. As the result, the normal and anomalous data cannot be accurately distinguished. To deal with the above-mentioned problems, multi-granularity contrastive learning for multivariate time series anomaly detection (MulGad) is proposed by combining contextual contrastive learning and inter-series contrastive learning. Specifically, the lambda-contextual contrastive learning (lambda-CCL) method based on time blocks is utilized to extract the fine-grained information within time series. In lambda-CCL, the learnable scale factor lambda n is used to adaptively adjust the penalty degree assigned to negative pair. The quality of contextual representations can thus be effectively improved. The cross-reconstruction based contrastive learning is designed to acquire the coarse-grained discriminative information among the augmented series. The cross-reconstruction is added to inter-series contrastive learning as an additional regularization. Hence, the robust representations are obtained, and the representation collapse problem is avoided. Moreover, a linear attention network is utilized to learn the temporal dependencies within the augmented series for the first time. The computational complexity of MulGad can thus be reduced. To validate the effectiveness of the proposed MulGad, it is compared with the eleven pertinent methods on the seven data sets. The outcomes demonstrate that MulGad is superior to the other eleven approaches. The code for MulGad is available at https://***/Xiaobowen-net/MulGad.

关键词： time series anomaly detection Contrastive learning Multi-granularity Cross-reconstruction

来源：评论

学校读者我要写书评

暂无评论

Applying quantum autoencoders for time series anomaly detection

引用

QUANTUM MACHINE INTELLIGENCE 2025年第1期7卷 1-21页

作者： Frehner, Robin Stockinger, Kurt Zurich Univ Appl Sci Zurich Switzerland

anomaly detection is an important problem with applications in various domains such as fraud detection, pattern recognition, or medical diagnosis. Several algorithms have been introduced using classical computing approaches. However, using quantum computing for solving anomaly detection problems in time series data is a widely unexplored research field. This paper explores the application of quantum autoencoders to time series anomaly detection. We investigate two primary techniques for classifying anomalies: (1) Analyzing the reconstruction error generated by the quantum autoencoder and (2) latent representation analysis. Our simulated experimental results, conducted across various ansaetze, demonstrate that quantum autoencoders consistently outperform classical deep learning-based autoencoders across multiple datasets. Specifically, quantum autoencoders achieve superior anomaly detection performance while utilizing 60-230 times fewer parameters and requiring five times fewer training iterations. In addition, we implement our quantum encoder on real quantum hardware. Our experimental results demonstrate that quantum autoencoders achieve anomaly detection performance on par with their simulated counterparts.

关键词： Quantum computing Quantum machine learning Quantum autoencoder time series anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Reevaluating the Potential of a Vanilla Transformer Encoder for Unsupervised time series anomaly detection in Sensor Applications

引用

SENSORS 2025年第8期25卷 2510-2510页

作者： Han, Chan Sik Kim, Hyungwon Lee, Keon Myung Chungbuk Natl Univ Dept Comp Sci Cheongju 28644 South Korea Chungbuk Natl Univ Dept Elect Engn Cheongju 28644 South Korea

Sensors generate extensive time series data across various domains, and effective methods for detecting anomalies in such data are still in high demand. Unsupervised time series anomaly detection provides practical approaches to addressing the challenges of collecting anomalous data. For effective anomaly detection, a range of deep-learning-based models have been explored to handle temporal patterns inherent in time series data. In particular, Transformer encoders have gained significant attention due to their ability to efficiently capture temporal dependencies. Various studies have attempted the architectural improvements of Transformer encoders to address the inherent complexity of time series data analysis. Unlike the previous studies, this work demonstrates that a vanilla Transformer encoder-based framework remains yet a competitive model for time series anomaly detection. Instead of architectural modification of the Transformer encoder, we identify key design choices and propose an asymmetric autoencoder-based framework incorporating those design choices with a vanilla Transformer encoder and a linear layer decoder. The proposed framework has been evaluated on a range of unsupervised time series anomaly detection benchmarks, and the experimental results show that it achieves performance that is either superior or competitive compared to state-of-the-art models.

关键词： time series anomaly detection unsupervised learning vanilla transformer encoders careful design choices

来源：评论

学校读者我要写书评

暂无评论

FEFM: Feature Extraction and Fusion Module for Enhanced time series anomaly detection 25

FEFM: Feature Extraction and Fusion Module for Enhanced Time...

引用

Proceedings of the 40th ACM/SIGAPP Symposium on Applied Computing

作者： Seong Hyun Jeon Keon Kim Yong Suk Choi Hanyang University Seoul Republic of Korea

ISBN: (纸本)9798400706295

time series data are utilized across various fields, including finance, healthcare, and manufacturing, where system reliability is crucial. Accordingly, extensive research on time series anomaly detection has been conducted. However, this task presents significant challenges due to high dimensionality, temporal dependencies, and noise in data. These challenges highlight the importance of effective feature extraction to capture meaningful data representations. In this study, we propose a novel Feature Extraction and Fusion Module (FEFM) specifically designed to enhance anomaly detection performance by extracting and fusing dimensional and temporal features of data. FEFM consists of three parts: Multi-Dimensional Feature Extractor (MDFE), Temporal Feature Extractor (TFE), and Feature Fusion Layers (FFL). MDFE and TFE extract dimensionally-driven and temporally-driven features from complex time series data. FFL then fuses these features with the original input, enabling the model to comprehensively understand the complex patterns. We evaluate the effectiveness of our method using seven benchmark datasets and two evaluation strategies, comparing its performance with other state-of-the-art methods. Experimental results demonstrate that our method outperforms others on various datasets, especially in reducing false positives. These results indicate that our method effectively extracts and fuses meaningful features from data, improving the reliability of anomaly detection systems.

关键词： time series anomaly detection

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：