Classical versus deep learning methods for anomaly detection in ECG using wavelet transformation

Abstract Time series anomaly detection has always been an important research direction. The early time series anomaly detection methods are mainly statistical methods and machine learning methods. With the powerful functions of deep neural network being continuously mined by researchers, the effect of deep neural network in anomaly detection task has been significantly better than the traditional methods. In view of the continuous development and application of deep neural networks such as transformer and graph neural network (GNN) in time series anomaly detection in recent years, the body of research lacks a comparative evaluation of deep learning methods in recent years. This paper studies various deep neural networks suitable for time series, which are divided into three categories according to anomaly detection methods. The evaluation is conducted on public datasets. By analyzing the evaluation criteria, this paper discusses the performance of each model, as well as the problems and development direction in the field of time series anomaly detection in the future. This study found that in the time series anomaly detection task, transformer is suitable for dealing with long-time series prediction, and studying the graph structure of time series may be the best way to deal with time series anomaly detection in the future

Download Full-text

Fault diagnosis systems for rotating machines operating with fluid-film bearings

Proceedings of the Institution of Mechanical Engineers Part J Journal of Engineering Tribology ◽

10.1177/13506501211073242 ◽

2022 ◽

pp. 135065012110732

Author(s):

Ivan Stebakov ◽

Alexey Kornaev ◽

Sergey Popov ◽

Leonid Savin

Keyword(s):

Deep Learning ◽

Fault Diagnosis ◽

Anomaly Detection ◽

Multilayer Perceptron ◽

Multiclass Classification ◽

Detection Accuracy ◽

Network Architectures ◽

Rotating Machines ◽

Learning Methods ◽

Deep Network

The paper deals with the application of deep learning methods to rotating machines fault diagnosis. The main challenge is to design a fault diagnosis system connected with multisensory measurement system that will be sensitive and accurate enough in detecting weak changes in rotating machines. The experimental part of the research presents the test rig and results of high-speed multisensory measurements. Six states of a rotating machine, including a normal one and five states with loosened mounting bolts and small unbalancing of the shaft, are under study. The application of deep network architectures including multilayer perceptron, convolutional neural networks, residual networks, autoencoders and their combination was estimated. The deep learning methods allowed to identify the most informative sensors, then solve the anomaly detection and the multiclass classification problems. An autoencoder based on ResNet architecture demonstrated the best result in anomaly detection. The accuracy of the proposed network is up to 100% while the accuracy of an expert is up to 65%. A one-dimensional convolutional neural network combined with a multilayer perceptron that contains a pretrained encoder demonstrated the best result in multiclass classification. The detailed fault detection accuracy with the determination of the specific fault is 83.3%. The combinations of known deep network architectures and application of the proposed approach of pretraining of the encoders together with using a block of inputs for one prediction demonstrated high efficiency.

Download Full-text

Robustness of Autoencoders for Anomaly Detection Under Adversarial Impact

Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence ◽

10.24963/ijcai.2020/173 ◽

2020 ◽

Author(s):

Adam Goodge ◽

Bryan Hooi ◽

See Kiong Ng ◽

Wee Siong Ng

Keyword(s):

Deep Learning ◽

Anomaly Detection ◽

Defense Mechanisms ◽

Model Performance ◽

Real Data ◽

Learning Methods ◽

Small Perturbations ◽

Unsupervised Anomaly Detection ◽

Latent Representations ◽

Supervised Image Classification

Detecting anomalies is an important task in a wide variety of applications and domains. Deep learning methods have achieved state-of-the-art performance in anomaly detection in recent years; unsupervised methods being particularly popular. However, deep learning methods can be fragile to small perturbations in the input data. This can be exploited by an adversary to deliberately hinder model performance; an adversarial attack. This phenomena has been widely studied in the context of supervised image classification since its discovery, however such studies for an anomaly detection setting are sorely lacking. Moreover, the plethora of defense mechanisms that have been proposed are often not applicable to unsupervised anomaly detection models. In this work, we study the effect of adversarial attacks on the performance of anomaly-detecting autoencoders using real data from a Cyber physical system (CPS) testbed with intervals of controlled, physical attacks as anomalies. An adversary would attempt to disguise these points as normal through adversarial perturbations. To combat this, we propose the Approximate Projection Autoencoder (APAE), which incorporates two defenses against such attacks into a general autoencoder. One of these involves a novel technique to improve robustness under adversarial impact by optimising latent representations for better reconstruction outputs.

Download Full-text