Anomaly Detection over Noisy Data using Learned Probability Distributions

By Eleazar Eskin. International Conference on Machine Learning, 2000

585 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Networks and Communications, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/580396c680951b7ff93defcac7cfe085dfe1814e

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

In Control and Systems Engineering

Citations in the corpus

Adversarially Learned One-Class Classifier for Novelty Detection (2018, background): “For instance, after extracting features from each sample, a distribution function is ﬁt on the samples from the target class, and samples far from this distribution are considered as outliers or novelty ( e.g ., [12, 51, 27]).”
Anomaly Detection in Streams with Extreme Value Theory (2017, methodology): “In [16], a more general mixture model is presented by Eskin, based on a majority distribution and an anomaly one.”
Semi-supervised detection of collective anomalies with an application in high energy particle physics (2012, methodology): “For example, Eskin [10] employs a similar mixture model for unsupervised anomaly detection under the assumption that anomalies can be modeled using a uniform distribution with a small mixture proportion, while Lauer [11] fits a Gaussian mixture model to an unlabeled data set by setting the anomalous”
Anomaly detection: A survey (2009, methodology): “While parametric techniques assume the knowledge of underlying distribution and estimate the parameters from the given data [Eskin 2000], non-parametric techniques do not generally assume knowledge of underlying distribution [Desforges et al.”
Outlier Analysis (2013, background): “mixture model specially as an anomaly component [154].”
Unsupervised Anomaly Detection in Network Intrusion Detection Using Clusters (2005, methodology): “In (Eskin 2000), a mixture model for explaining the presence of anomalies is presented, and machine learning techniques are used to estimate the probability distributions of the mixture to detect anomalies.”
Robust Anomaly Detection Using Support Vector Machines (2003, background): “If an attack occurred during the training process, the undesired, intrusive behavior could get established as a part of the anomaly detector’s model of normal behavior and thus undermine its ability to detect subsequent occurrences [1] [7] [8].”
Progress in Outlier Detection Techniques: A Survey (2019, unknown): “The latter method does not have any assumption of prior knowledge of the distributionmodel [98].”
Generative Probabilistic Novelty Detection with Adversarial Autoencoders (2018, methodology): “Novelty detection methods can be statistical and probabilistic based [15, 16], distance based [17], and also based on self-representation [8].”
A survey of network anomaly detection techniques (2016, background): “The authors of Eskin (2000) assumed from an intrusion detection perspective that the set of system calls with the probability of 1 λ is a legitimate use of the system and the intrusions have the probability of λ.”
Robust Estimation in Signal Processing: A Tutorial-Style Treatment of Fundamental Concepts (2012, background): “The first approach consists of learning the probability distributions of the data, which are modeled as a mixture of outliers and clean data [20].”
Outlier Detection (2008, background): “In this sense, depth functions can measure the “extremeness” or “outlyingness” of a data point with respect to a given data set.”
Anomaly pattern detection in categorical datasets (2008, background): “This problem has received considerable attention, particularly in the ﬁelds of network intrusion detection [14, 8, 6] and the monitoring of public health data for disease outbreak detection [15, 13].”
Feature bagging for outlier detection (2005, unknown): “In statistical techniques [5, 7, 12], the data points are typically modeled using a stochastic distribution, and points are determined to be outliers depending on their relationship with this model.”
Unsupervised Class-Incremental Learning Through Confusion (2021, background): “Current approaches can be largely categorized by statistical, distance-based, and deep learning methods (Eskin (2000); Yamanishi et al. (2004); Knorr et al. (2000); Hautamaki et al. (2004); Sabokrou et al. (2018); Kliger & Fleishman (2018)). Recent techniques involve using a threshold to determine c”
An immunity-based technique to characterize intrusions in computer networks (2002, methodology): “There are multiple approaches to such anomaly detection [5], [7], [13]–[15], [22], [23], [26] and most of them work by building a model or profile of the system that reflects its normal behavior.”