Results 11  20
of
343
Graphbased anomaly detection
 In KDD
, 2003
"... Anomaly detection is an area that has received much attention in recent years. It has a wide variety of applications, including fraud detection and network intrusion detection. A good deal of research has been performed in this area, often using strings or attributevalue data as the medium from whi ..."
Abstract

Cited by 106 (2 self)
 Add to MetaCart
(Show Context)
Anomaly detection is an area that has received much attention in recent years. It has a wide variety of applications, including fraud detection and network intrusion detection. A good deal of research has been performed in this area, often using strings or attributevalue data as the medium from which anomalies are to be extracted. Little work, however, has focused on anomaly detection in graphbased data. In this paper, we introduce two techniques for graphbased anomaly detection. In addition, we introduce methods for calculating the regularity of a graph, with applications to anomaly detection. We hypothesize that these methods will prove useful both for finding anomalies, and for determining the likelihood of successful anomaly detection within graphbased data. We provide experimental results using both realworld network intrusion data and artificiallycreated data. 1.
Comparing Dynamic Causal Models
 NEUROIMAGE
, 2004
"... This article describes the use of Bayes factors for comparing Dynamic Causal Models (DCMs). DCMs are used to make inferences about effective connectivity from functional Magnetic Resonance Imaging (fMRI) data. These inferences, however, are contingent upon assumptions about model structure, that is, ..."
Abstract

Cited by 105 (35 self)
 Add to MetaCart
(Show Context)
This article describes the use of Bayes factors for comparing Dynamic Causal Models (DCMs). DCMs are used to make inferences about effective connectivity from functional Magnetic Resonance Imaging (fMRI) data. These inferences, however, are contingent upon assumptions about model structure, that is, the connectivity pattern between the regions included in the model. Given the current lack of detailed knowledge on anatomical connectivity in the human brain, there are often considerable degrees of freedom when defining the connectional structure of DCMs. In addition, many plausible scientific hypotheses may exist about which connections are changed by experimental manipulation, and a formal procedure for directly comparing these competing hypotheses is highly desirable. In this article, we show how Bayes factors can be used to guide choices about model structure, both with regard to the intrinsic connectivity pattern and the contextual modulation of individual connections. The combined use of Bayes factors and DCM thus allows one to evaluate competing scientific theories about the architecture of largescale neural networks and the neuronal interactions that mediate perception and cognition.
Unsupervised models for morpheme segmentation and morphology learning
 ACM Trans. Speech Lang. Process
, 2007
"... We present a model family called Morfessor for the unsupervised induction of a simple morphology from raw text data. The model is formulated in a probabilistic maximum a posteriori framework. Morfessor can handle highly inflecting and compounding languages where words can consist of lengthy sequence ..."
Abstract

Cited by 101 (7 self)
 Add to MetaCart
(Show Context)
We present a model family called Morfessor for the unsupervised induction of a simple morphology from raw text data. The model is formulated in a probabilistic maximum a posteriori framework. Morfessor can handle highly inflecting and compounding languages where words can consist of lengthy sequences of morphemes. A lexicon of word segments, called morphs, is induced from the data. The lexicon stores information about both the usage and form of the morphs. Several instances of the model are evaluated quantitatively in a morpheme segmentation task on different sized sets of Finnish as well as English data. Morfessor is shown to perform very well compared to a widely known benchmark algorithm, in particular on Finnish data.
Bayesian Optimization Algorithm: From Single Level to Hierarchy
, 2002
"... There are four primary goals of this dissertation. First, design a competent optimization algorithm capable of learning and exploiting appropriate problem decomposition by sampling and evaluating candidate solutions. Second, extend the proposed algorithm to enable the use of hierarchical decompositi ..."
Abstract

Cited by 99 (18 self)
 Add to MetaCart
(Show Context)
There are four primary goals of this dissertation. First, design a competent optimization algorithm capable of learning and exploiting appropriate problem decomposition by sampling and evaluating candidate solutions. Second, extend the proposed algorithm to enable the use of hierarchical decomposition as opposed to decomposition on only a single level. Third, design a class of difficult hierarchical problems that can be used to test the algorithms that attempt to exploit hierarchical decomposition. Fourth, test the developed algorithms on the designed class of problems and several realworld applications. The dissertation proposes the Bayesian optimization algorithm (BOA), which uses Bayesian networks to model the promising solutions found so far and sample new candidate solutions. BOA is theoretically and empirically shown to be capable of both learning a proper decomposition of the problem and exploiting the learned decomposition to ensure robust and scalable search for the optimum across a wide range of problems. The dissertation then identifies important features that must be incorporated into the basic BOA to solve problems that are not decomposable on a single level, but that can still be solved by decomposition over multiple levels of difficulty. Hierarchical
Akaike’s information criterion and recent developments in information complexity
 Journal of Mathematical Psychology
"... criterion (AIC). Then, we present some recent developments on a new entropic or information complexity (ICOMP) criterion of Bozdogan (1988a, 1988b, 1990, 1994d, 1996, 1998a, 1998b) for model selection. A rationale for ICOMP as a model selection criterion is that it combines a badnessoffit term (su ..."
Abstract

Cited by 99 (9 self)
 Add to MetaCart
(Show Context)
criterion (AIC). Then, we present some recent developments on a new entropic or information complexity (ICOMP) criterion of Bozdogan (1988a, 1988b, 1990, 1994d, 1996, 1998a, 1998b) for model selection. A rationale for ICOMP as a model selection criterion is that it combines a badnessoffit term (such as minus twice the maximum log likelihood) with a measure of complexity of a model differently than AIC, or its variants, by taking into account the interdependencies of the parameter estimates as well as the dependencies of the model residuals. We operationalize the general form of ICOMP based on the quantification of the concept of overall model complexity in terms of the estimated inverseFisher information matrix. This approach results in an approximation to the sum of two KullbackLeibler distances. Using the correlational form of the complexity, we further provide yet another form of ICOMP to take into account the interdependencies (i.e., correlations) among the parameter estimates of the model. Later, we illustrate the practical utility and the importance of this new model selection criterion by providing several
A tutorial introduction to the minimum description length principle
 in Advances in Minimum Description Length: Theory and Applications. 2005
"... ..."
(Show Context)
Reconciling simplicity and likelihood principles in perceptual organization
 Psychological Review
, 1996
"... Two principles of perceptual organization have been proposed. The likelihood principle, following H. L. E yon Helmholtz ( 1910 / 1962), proposes that perceptual organization is chosen to correspond to the most likely distal layout. The simplicity principle, following Gestalt psychology, suggests tha ..."
Abstract

Cited by 86 (17 self)
 Add to MetaCart
(Show Context)
Two principles of perceptual organization have been proposed. The likelihood principle, following H. L. E yon Helmholtz ( 1910 / 1962), proposes that perceptual organization is chosen to correspond to the most likely distal layout. The simplicity principle, following Gestalt psychology, suggests that perceptual organization is chosen to be as simple as possible. The debate between these two views has been a central topic in the study of perceptual organization. Drawing on mathematical results in A. N. Kolmogorov's ( 1965)complexity heory, the author argues that simplicity and likelihood are not in competition, but are identical. Various implications for the theory of perceptual organization and psychology more generally are outlined. How does the perceptual system derive a complex and structured description of the perceptual world from patterns of activity at the sensory receptors? Two apparently competing theories of perceptual organization have been influential. The first, initiated by Helmholtz ( 1910/1962), advocates the likelihood principle: Sensory input will be organized into the most probable distal object or event consistent with that input. The second, initiated by Wertheimer and developed by other Gestalt psychologists, advocates what Pomerantz and Kubovy (1986) called the simplicity principle: The perceptual system is viewed as finding the simplest, rather than the most likely, perceptual organization consistent with the sensory input '. There has been considerable theoretical nd empirical controversy concerning whether likelihood or simplicity is the governing principle of perceptual organization (e.g., Hatfield, &
Unsupervised Discovery of Morphemes
, 2002
"... We present two methods for unsupervised segmentation of words into morphemelike units. The model utilized is especially suited for languages with a rich morphology, such as Finnish. The first method is based on the Minimum Description Length (MDL) principle and works online. In the second met ..."
Abstract

Cited by 80 (16 self)
 Add to MetaCart
(Show Context)
We present two methods for unsupervised segmentation of words into morphemelike units. The model utilized is especially suited for languages with a rich morphology, such as Finnish. The first method is based on the Minimum Description Length (MDL) principle and works online. In the second method, Maximum Likelihood (ML) optimization is used. The quality of the segmentations is measured using an evaluation method that compares the segmentations produced to an existing morphological analysis. Experiments on both Finnish and English corpora show that the presented methods perform well compared to a current stateoftheart system.
Beyond tracking: modelling activity and understanding behaviour
 International Journal of Computer Vision
, 2006
"... In this work, we present a unified bottomup and topdown automatic model selection based approach for modelling complex activities of multiple objects in cluttered scenes. An activity of multiple objects is represented based on discrete scene events and their behaviours are modelled by reasoning ab ..."
Abstract

Cited by 78 (14 self)
 Add to MetaCart
(Show Context)
In this work, we present a unified bottomup and topdown automatic model selection based approach for modelling complex activities of multiple objects in cluttered scenes. An activity of multiple objects is represented based on discrete scene events and their behaviours are modelled by reasoning about the temporal and causal correlations among different events. This is significantly different from the majority of the existing techniques that are centred on object tracking followed by trajectory matching. In our approach, objectindependent events are detected and classified by unsupervised clustering using ExpectationMaximisation (EM) and classified using automatic model selection based on Schwarz’s Bayesian Information Criterion (BIC). Dynamic Probabilistic Networks (DPNs) are formulated for modelling the temporal and causal correlations among discrete events for robust and holistic scenelevel behaviour interpretation. In particular, we developed a Dynamically MultiLinked Hidden Markov Model (DMLHMM) based on the discovery of salient dynamic interlinks among multiple temporal processes corresponding to multiple event classes. A DMLHMM is built using BIC based factorisation resulting in its topology being intrinsically determined by the underlying causality and temporal order among events. Extensive experiments are conducted on modelling activities captured in different indoor and