Anomaly Detection using One-Class Neural Networks

By Raghavendra Chalapathy, Aditya Krishna Menon, Sanjay Chawla. 2018

422 citations recorded by the source corpus.

We propose a one-class neural network (OC-NN) model to detect anomalies in complex data sets. OC-NN combines the ability of deep networks to extract a progressively rich representation of data with the one-class objective of creating a tight envelope around normal data. The OC-NN approach breaks new ground for the following crucial reason: data representation in the hidden layer is driven by the OC-NN objective and is thus customized for anomaly detection. This is a departure from other approaches which use a hybrid approach of learning deep features using an autoencoder and then feeding the features into a separate anomaly detection method like one-class SVM (OC-SVM). The hybrid OC-SVM approach is sub-optimal because it is unable to influence representational learning in the hidden layers. A comprehensive set of experiments demonstrate that on complex data sets (like CIFAR and GTSRB), OC-NN performs on par with state-of-the-art methods and outperformed conventional shallow methods in some scenarios.

Subfields: Artificial Intelligence, Computer Networks and Communications

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 Computer Networks and Communications

Citations in the corpus

SentiNet: Detecting Localized Universal Attacks Against Deep Learning Systems (2018, background): “ntage of the richer data provided by the neuron outputs during the inference process. Works have shown techniques to detect anomalous high dimensional data using one class neural networks (see, e.g., [7], [16]) which could enhance our current framework. We hope SentiNet inspires further approaches t”
SSD: A Unified Framework for Self-Supervised Outlier Detection (2021, methodology): “…on developing methods particularly for single-class modeling in anomaly detection (Perera et al., 2019; Schlegl et al., 2017; Ruff et al., 2018; Chalapathy et al., 2018b; Golan & El-Yaniv, 2018; Wang et al., 2019), some recent work achieve success in both OOD detection and anomaly detection…”
HRN: A Holistic Approach to One Class Learning (2020, background): “For a survey of deep learning based one-class anomaly detection methods, see [8].”
SentiNet: Detecting Physical Attacks Against Deep Learning Systems (2018, background): “Works have shown techniques to detect anomalous high dimensional data using one class neural networks (see, e.g., [7], [16]) which could enhance our current framework.”
Deep Learning for Anomaly Detection (2020, background): “2 OC-NN [27] Keras https://git.”
Deep Semi-Supervised Anomaly Detection (2019, methodology): “Semi-supervised anomaly detection setup MNIST, Fashion-MNIST, and CIFAR-10 all have ten classes from which we derive ten AD setups on each dataset following previous works (Ruff et al., 2018; Chalapathy et al., 2018; Golan & El-Yaniv, 2018).”
Learning Deep Features for One-Class Classification (2018, methodology): “Method proposed in [5] applied on the extracted features from the AlexNet and VGG16 networks.”
Student-Teacher Feature Pyramid Matching for Anomaly Detection (2021, background): “Such deep features prove very effective in capturing the intrinsic characteristics of the normal data manifold [3, 10, 24, 32, 47].”
DRÆM – A discriminatively trained reconstruction embedding for surface anomaly detection (2021, background): “These are based on one-class methods [22, 7] which attempt to estimate a decision boundary around anomaly-free data that separates it from anomalous samples by assuming a unimodal distribution of the anomaly-free data.”
Detecting AI Trojans Using Meta Neural Analysis (2019, methodology): “Therefore, we leverage the approach in [11] which generalizes the one-class optimization goal to neural networks by replacing the term || w || 2 in Eqn.6 with the Frobenius norm [1] of all the parameters in the network.”
VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization (2021, positive): “Some of the works handled these two steps independently [25], [26], while others achieved better results by doing the two steps collectively, in order to extract the best features for subsequent anomaly detection [27], [28].”
A Unifying Review of Deep and Shallow Anomaly Detection (2020, neutral): “Deep SVDD [137], [144], [145], [327] and deep OC-SVM variants [136], [328] employ a hypersphere model fθ(x) = ‖φω(x) − c‖2 −R2 and linear model fθ(x) = ρ − 〈φω(x),w〉 with explicit neural feature maps φω(·) in (16) and (20), respectively.”
You Are What You Do: Hunting Stealthy Malware via Data Provenance Analysis (2020, methodology): “More complex machine learning techniques, such as LSTM [58], Tree-structured LSTM [98], Graph Convolu-tional Networks [64], and One-class Neural Networks [36], [88] could possibly further improve the detection accuracy, yet they may also introduce a higher cost.”
Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings (2019, background): “Recently, efforts have been made to improve anomaly detection for one-class or multi-class classification [2, 3, 10, 11, 21, 28, 29].”
Memorizing Normality to Detect Anomaly: Memory-Augmented Deep Autoencoder for Unsupervised Anomaly Detection (2019, methodology): “A natural choice for handling the problem is thus the oneclass classification methods, such as one-class SVM [5, 34] and deep one-class networks [31, 3], which seeks to learn a discriminative hyperplane surrounding the normal samples.”
The MVTec Anomaly Detection Dataset: A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection (2021, background): “A popular approach is to select an arbitrary subset of classes, re-label them as outliers, and train a novelty detection system solely on the remaining inlier classes (An and Cho 2015; Chalapathy et al. 2018; Ruff et al. 2018; Burlina et al. 2019).”
MVTec AD — A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection (2019, background): “A popular approach [1, 7, 21] is to select an arbitrary subset of classes, re-label them as outliers, and train a novelty detection system solely on the remaining inlier classes.”