Results 1  10
of
917
Dynamic Bayesian Networks: Representation, Inference and Learning
, 2002
"... Modelling sequential data is important in many areas of science and engineering. Hidden Markov models (HMMs) and Kalman filter models (KFMs) are popular for this because they are simple and flexible. For example, HMMs have been used for speech recognition and biosequence analysis, and KFMs have bee ..."
Abstract

Cited by 564 (3 self)
 Add to MetaCart
Modelling sequential data is important in many areas of science and engineering. Hidden Markov models (HMMs) and Kalman filter models (KFMs) are popular for this because they are simple and flexible. For example, HMMs have been used for speech recognition and biosequence analysis, and KFMs have been used for problems ranging from tracking planes and missiles to predicting the economy. However, HMMs
and KFMs are limited in their “expressive power”. Dynamic Bayesian Networks (DBNs) generalize HMMs by allowing the state space to be represented in factored form, instead of as a single discrete random variable. DBNs generalize KFMs by allowing arbitrary probability distributions, not just (unimodal) linearGaussian. In this thesis, I will discuss how to represent many different kinds of models as DBNs, how to perform exact and approximate inference in DBNs, and how to learn DBN models from sequential data.
In particular, the main novel technical contributions of this thesis are as follows: a way of representing
Hierarchical HMMs as DBNs, which enables inference to be done in O(T) time instead of O(T 3), where T is the length of the sequence; an exact smoothing algorithm that takes O(log T) space instead of O(T); a simple way of using the junction tree algorithm for online inference in DBNs; new complexity bounds on exact online inference in DBNs; a new deterministic approximate inference algorithm called factored frontier; an analysis of the relationship between the BK algorithm and loopy belief propagation; a way of
applying RaoBlackwellised particle filtering to DBNs in general, and the SLAM (simultaneous localization
and mapping) problem in particular; a way of extending the structural EM algorithm to DBNs; and a variety of different applications of DBNs. However, perhaps the main value of the thesis is its catholic presentation of the field of sequential data modelling.
The Bayes Net Toolbox for MATLAB
 Computing Science and Statistics
, 2001
"... The Bayes Net Toolbox (BNT) is an opensource Matlab package for directed graphical models. BNT supports many kinds of nodes (probability distributions), exact and approximate inference, parameter and structure learning, and static and dynamic models. BNT is widely used in teaching and research: the ..."
Abstract

Cited by 176 (2 self)
 Add to MetaCart
The Bayes Net Toolbox (BNT) is an opensource Matlab package for directed graphical models. BNT supports many kinds of nodes (probability distributions), exact and approximate inference, parameter and structure learning, and static and dynamic models. BNT is widely used in teaching and research: the web page has received over 28,000 hits since May 2000. In this paper, we discuss a broad spectrum of issues related to graphical models (directed and undirected), and describe, at a highlevel, how BNT was designed to cope with them all. We also compare BNT to other software packages for graphical models, and to the nascent OpenBayes effort.
Inferring Subnetworks from Perturbed Expression Profiles
, 2001
"... Genomewide expression profiles of genetic mutants provide a wide variety of measurements of cellular responses to perturbations. Typical analysis of such data identifies genes affected by perturbation and uses clustering to group genes of similar function. In this paper we discover a finer structur ..."
Abstract

Cited by 159 (12 self)
 Add to MetaCart
Genomewide expression profiles of genetic mutants provide a wide variety of measurements of cellular responses to perturbations. Typical analysis of such data identifies genes affected by perturbation and uses clustering to group genes of similar function. In this paper we discover a finer structure of interactions between genes, such as causality, mediation, activation, and inhibition by using a Bayesian network framework. We extend this framework to correctly handle perturbations, and to identify significant subnetworks of interacting genes. We apply this method to expression data of S. cerevisiae mutants and uncover a variety of structured metabolic, signaling and regulatory pathways. Contact: danab@cs.huji.ac.il
Mediation in experimental and nonexperimental studies: new procedures and recommendations
 Psychological Methods
, 2002
"... Mediation is said to occur when a causal effect of some variable X on an outcome Y is explained by some intervening variable M. The authors recommend that with small to moderate samples, bootstrap methods (B. Efron & R. Tibshirani, 1993) be used to assess mediation. Bootstrap tests are powerful beca ..."
Abstract

Cited by 126 (2 self)
 Add to MetaCart
Mediation is said to occur when a causal effect of some variable X on an outcome Y is explained by some intervening variable M. The authors recommend that with small to moderate samples, bootstrap methods (B. Efron & R. Tibshirani, 1993) be used to assess mediation. Bootstrap tests are powerful because they detect that the sampling distribution of the mediated effect is skewed away from 0. They argue that R. M. Baron and D. A. Kenny’s (1986) recommendation of first testing the X → Y association for statistical significance should not be a requirement when there is a priori belief that the effect size is small or suppression is a possibility. Empirical examples and computer setups for bootstrap analyses are provided. Mediation models of psychological processes are popular because they allow interesting associations to be decomposed into components that reveal possible causal mechanisms. These models are useful for theory development and testing as well as for the identification of possible points of intervention in applied work. Mediation is equally of interest to experimental psychologists as it is to those who study naturally occurring processes through nonexperimental studies. For example, social–cognitive psychologists are interested in showing that the effects of cognitive priming on attitude change are mediated by the accessibility of certain beliefs (Eagly & Chaiken, 1993). Developmental psychologists use longitudinal methods to study how parental unemployment can have adverse effects on child behavior through its intervening effect on quality of parenting (Conger et al., 1990). Mediation analysis is also used in organizational
Causes and explanations: A structuralmodel approach
 In Proceedings IJCAI01
, 2001
"... We propose a new definition of actual causes, using structural equations to model counterfactuals. We show that the definition yields a plausible and elegant account of causation that handles well examples which have caused problems for other definitions ..."
Abstract

Cited by 118 (9 self)
 Add to MetaCart
We propose a new definition of actual causes, using structural equations to model counterfactuals. We show that the definition yields a plausible and elegant account of causation that handles well examples which have caused problems for other definitions
The Nature of Theory in Information Systems
 MIS Quarterly
, 2006
"... The aim of this research essay is to examine the structural nature of theory in information systems. Despite the importance of theory, questions relating to its form and structure are neglected in comparison with questions relating to epistemology. The essay addresses issues of causality, explanatio ..."
Abstract

Cited by 101 (2 self)
 Add to MetaCart
The aim of this research essay is to examine the structural nature of theory in information systems. Despite the importance of theory, questions relating to its form and structure are neglected in comparison with questions relating to epistemology. The essay addresses issues of causality, explanation, prediction and generalization that underlie an understanding of theory. A taxonomy is proposed that classifies information systems theories with respect to the manner in which four central goals are addressed: analysis, explanation, prediction and prescription. Five interrelated types of theory are distinguished: (i) theory for analysing; (ii) theory for explaining, (iii) theory for predicting; (iv) theory for explaining and predicting; and (v) theory for design and action. Examples illustrate the nature of each theory type. The applicability of the taxonomy is demonstrated by classifying a sample of journal articles. The paper contributes by showing that multiple views of theory exist and by exposing the assumptions underlying different viewpoints. In addition, it is suggested that the type of theory under development can influence the choice of an epistemological approach. Support is given for the legitimacy and value of each theory type. The building of integrated bodies of theory that encompass all theory types is advocated.
indexing, clustering, and retrieving system history
, 2005
"... In operating today’s complex systems, the lack of a systematic way to capture and query the essential system state characterizing an incident of performance failure or unavailability makes it difficult for operators to distinguish recurring problems from new ones, to leverage previous diagnostic eff ..."
Abstract

Cited by 91 (7 self)
 Add to MetaCart
In operating today’s complex systems, the lack of a systematic way to capture and query the essential system state characterizing an incident of performance failure or unavailability makes it difficult for operators to distinguish recurring problems from new ones, to leverage previous diagnostic efforts, or to establish whether problems seen at different installations of the same site are similar or distinct. We present a method for automatically extracting from a running system an indexable signature that distills the essential system state associated with a problem and can be subjected to clustering and similaritybased retrieval to identify when an observed system state is similar to a previouslyobserved state. This allows operators to identify and quantify the frequency of recurrent problems, or to leverage problem observations at one site in diagnosing problems at another site. We show that the naive approach to constructing these signatures is ineffective, leading us to a more sophisticated approach based on statistical modeling and inference. Our method requires only that the system’s metric of merit as well as a collection of lowerlevel operational metrics be collected, as is done by existing commercial monitoring tools. Even if the traces have no annotations of prior diagnoses of observed incidents (as is typical), our technique successfully clusters system states corresponding to similar problems, allowing diagnosticians to identify recurring problems or to name the “syndrome ” of a group of problems later. We validate our approach on both synthetic traces and several weeks of production traces from a customerfacing geoplexed 24 × 7 system; in the latter case, our approach identified a recurring problem that had required extensive manual diagnosis, and also aided the operators in correcting a previous misdiagnosis of a different problem. “Those who cannot remember the past are condemned to repeat it. ” — George Santayana 1
Lifted firstorder probabilistic inference
 In Proceedings of IJCAI05, 19th International Joint Conference on Artificial Intelligence
, 2005
"... Most probabilistic inference algorithms are specified and processed on a propositional level. In the last decade, many proposals for algorithms accepting firstorder specifications have been presented, but in the inference stage they still operate on a mostly propositional representation level. [Poo ..."
Abstract

Cited by 88 (7 self)
 Add to MetaCart
Most probabilistic inference algorithms are specified and processed on a propositional level. In the last decade, many proposals for algorithms accepting firstorder specifications have been presented, but in the inference stage they still operate on a mostly propositional representation level. [Poole, 2003] presented a method to perform inference directly on the firstorder level, but this method is limited to special cases. In this paper we present the first exact inference algorithm that operates directly on a firstorder level, and that can be applied to any firstorder model (specified in a language that generalizes undirected graphical models). Our experiments show superior performance in comparison with propositional exact inference. 1
Direct and Indirect Effects
, 2005
"... The direct effect of one event on another can be defined and measured by holding constant all intermediate variables between the two. Indirect effects present conceptual and practical difficulties (in nonlinear models), because they cannot be isolated by holding certain variables constant. This pape ..."
Abstract

Cited by 76 (22 self)
 Add to MetaCart
The direct effect of one event on another can be defined and measured by holding constant all intermediate variables between the two. Indirect effects present conceptual and practical difficulties (in nonlinear models), because they cannot be isolated by holding certain variables constant. This paper presents a new way of defining the effect transmitted through a restricted set of paths, without controlling variables on the remaining paths. This permits the assessment of a more natural type of direct and indirect effects, one that is applicable in both linear and nonlinear models and that has broader policyrelated interpretations. The paper establishes conditions under which such assessments can be estimated consistently from experimental and nonexperimental data, and thus extends pathanalytic techniques to nonlinear and nonparametric models.