Anomaly Detection Using Autoencoders with Nonlinear Dimensionality Reduction

By Mayu Sakurada, Takehisa Yairi. 2014

1,400 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Control and Systems Engineering

In Signal Processing

Citations in the corpus

A practical tutorial on autoencoders for nonlinear feature fusion: Taxonomy, models, software and guidelines (2018, methodology): “] these terms, along with deviants, discordants or abnormalities, refer to the same concept. The telemetry obtained from spacecrafts is quite complex, made up of hundreds of variables. The authors of [127] propose the use of basic and denoising AEs for facing anomaly detection taking advantage of th”
DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detection (2023, methodology): “Such methods can be dated back to AutoEncoder [58], LSTM-VAE [50], Spectral Residual (SR) [54], RobustTAD [26] and so on.”
Deep Semi-Supervised Anomaly Detection (2019, background): “For example autoencoding models [57, 26], which make up the predominant class of approaches to deep AD [24, 58, 4, 19, 72, 12, 9], can be understood as implicitly maximizing the mutual information I ( X ; Z ) via the reconstruction objective under some regularization of the latent code Z .”
Towards Total Recall in Industrial Anomaly Detection (2021, background): “Existing work on cold-start, industrial visual anomaly detection relies on learning a model of the nominal distribution via autoencoding methods [12, 36, 44], GANs [2, 39, 43], or other unsupervised adaptation methods [42, 56].”
Deep Generative Design: Integration of Topology Optimization and Generative Models (2019, methodology): “Reconstruction errors, which are the loss functions of autoencoder, are widely used to detect anomaly [50].”
Deep One-Class Classification (2018, methodology): “…2016; Erfani et al., 2016; Sabokrou et al., 2016), by plugging the learned embeddings into classical AD methods, but also in fully deep approaches, by directly employing the reconstruction error as an anomaly score (Hawkins et al., 2002; Sakurada & Yairi, 2014; An & Cho, 2015; Chen et al., 2017).”
Generative Probabilistic Novelty Detection with Adversarial Autoencoders (2018, background): “Similarly to previous work on novelty detection [38, 39, 40, 7, 11, 13], such networks are based on autoencoders [41, 42].”
Kitsune: An Ensemble of Autoencoders for Online Network Intrusion Detection (2018, background): “In [24], the authors proposed the generic use of an autoencoder for detecting anomalies.”
Deep Learning for Time Series Anomaly Detection: A Survey (2022, neutral): “Sakurada and Yairi [150] shows how AEs can be used for dimensionality reduction in MTS as a preprocessing step for anomaly detection.”
Learning Discriminative Reconstructions for Unsupervised Outlier Removal (2015, methodology): “We notice reconstruction error has also been used for novelty detection [10, 16, 20], which is a related but fundamentally different task with unsupervised outlier removal: the training data in novelty detection are all positive, while that in unsupervised outlier removal are quite noisy.”
Learning Graph Structures With Transformer for Multivariate Time-Series Anomaly Detection in IoT (2021, methodology): “[36] conducted dimentionality reduction based on autoencoders for anomaly detection.”
Deep Learning Anomaly Detection as Support Fraud Investigation in Brazilian Exports and Anti-Money Laundering (2016, methodology): “These results are supported by Sakurada and Yairi [26]: PCA is computationally more expensive than AutoEncoder because it “basically requires to hold all the training samples”.”
Unsupervised Deep Anomaly Detection for Multi-Sensor Time-Series Signals (2021, methodology): “2) To reduce the influence of noisy data, we improve Deep Convolutional Autoencoder with a Maximum Mean Discrepancy (MMD) penalty.”
Anomaly Detection for IoT Time-Series Data: A Survey (2020, methodology): “Reconstruction error is used to detect anomalies using a range of autoencoder methods over satellite telemetry and artiﬁcial data in [88].”
Sub-Image Anomaly Detection with Deep Pyramid Correspondences (2020, background): “Principal components have been extended to non-linear functions learned by autoencoders [27], including both denoising as well as variational autoencoders (VAEs).”