Hidden Markov models (HMMs) are widely used models for sequential data. As with other probabilisticmodels, they require the specification of local conditional probability distributions, which can be too difficult and...
详细信息
ISBN:
(纸本)9781479956180
Hidden Markov models (HMMs) are widely used models for sequential data. As with other probabilisticmodels, they require the specification of local conditional probability distributions, which can be too difficult and error-prone, especially when data are scarce or costly to acquire. The imprecise HMM (iHMM) generalizes HMMs by allowing the quantification to be done by sets of, instead of single, probability distributions. iHMMs have the ability to suspend judgment when there is not enough statistical evidence, and can serve as a sensitivity analysis tool for standard non-stationary HMMs. In this paper, we formalize iHMMs and develop efficient inference algorithms to address standard HMM usage such as the computation of likelihoods and most probable explanations. Experiments with real data show that iHMMs produce more reliable inferences without compromising efficiency.
We propose a framework for the selection of failure countermeasures and repair actions, based on Decision Networks (DN). We show, through specific examples, that standard probabilistic inference on DN can be used to c...
详细信息
ISBN:
(纸本)9781479928477
We propose a framework for the selection of failure countermeasures and repair actions, based on Decision Networks (DN). We show, through specific examples, that standard probabilistic inference on DN can be used to compute system reliability, component importance measures, as well as to select the best (in terms of Maximum Expected Utility) set of failure countermeasures to activate. Finally, by exploiting the DN formalism, both modeling and analysis capabilities are improved with respect to standard combinatorial models, without resorting to the complexity of global state-space models.
This paper develops a mathematical and computational framework for analyzing the expected performance of Bayesian data fusion, or joint statistical inference, within a sensor network. We use variational techniques to ...
详细信息
This paper develops a mathematical and computational framework for analyzing the expected performance of Bayesian data fusion, or joint statistical inference, within a sensor network. We use variational techniques to obtain the posterior expectation as the optimal fusion rule under a deterministic constraint and a quadratic cost, and study the smoothness and other properties of its classification performance. For a certain class of fusion problems, we prove that this fusion rule is also optimal in a much wider sense and satisfies strong asymptotic convergence results. We show how these results apply to a variety of examples with Gaussian, exponential and other statistics, and discuss computational methods for determining the fusion system's performance in more general, large-scale problems. These results are motivated by studying the performance of fusing multi-modal radar and acoustic sensors for detecting explosive substances, but have broad applicability to other Bayesian decision problems.
The primary focus of this work is to explore and describe a proposed decision process for use in controlling heating, ventilation, and air conditioning systems in buildings. The larger problem that houses the decision...
详细信息
The primary focus of this work is to explore and describe a proposed decision process for use in controlling heating, ventilation, and air conditioning systems in buildings. The larger problem that houses the decision process is structured as a multi-agent building control system. This is because multi-agent systems are an effective way to incorporate new sources of external and sporadic information into the basic system. In this study, sources of occupant information are used to demonstrate the decision process. The decision process begins with agents formulating context specific opinions using a method that is derived from Bayesian Networks and probabilisticgraphical modeling. From there, agents convey their preferences using conditional game theory, setting up the group of agents for action. While there are a variety of action procedures that could be used, the process was developed with Nash equilibriums in mind. The Nash equilibriums, however, were not yet successfully implemented due to the early stages of the social influence modeling. Instead, a basic social welfare function was employed to demonstrate the decision process for a setpoint decision context. A three zone, variable air volume simulator was developed in MATLAB to represent a physical environment and to act as a reference case for typical PI control. The simulator is then used for exploring and describing the decision process implementation, which is run in parallel to the basic system simulator. The Bayesian Network-based opinion formulation method developed by this work holds promise and has been found to merit further development. A component of the conditional game structure, rejectability, was not found to be a natural context for use in the setpoint decision context, but it may be improved upon. Overall, systems integration was desired and appears possible by implementing the decision process.
Critical to high-dimensional statistical estimation is to exploit the structure in the data distribution. probabilistic graphical models provide an efficient framework for representing complex joint distributions of r...
详细信息
Critical to high-dimensional statistical estimation is to exploit the structure in the data distribution. probabilistic graphical models provide an efficient framework for representing complex joint distributions of random variables through their conditional dependency graph, and can be adapted to many high-dimensional machine learning applications. This dissertation develops the probabilisticgraphical modeling technique for three statistical estimation problems arising in real-world applications: distributed and parallel learning in networks, missing-value prediction in recommender systems, and emerging topic detection in text corpora. The common theme behind all proposed methods is a combination of parsimonious representation of uncertainties in the data, optimization surrogate that leads to computationally efficient algorithms, and fundamental limits of estimation performance in high dimension. More specifically, the dissertation makes the following theoretical contributions: (1) We propose a distributed and parallel framework for learning the parameters in Gaussian graphicalmodels that is free of iterative global message passing. The proposed distributed estimator is shown to be asymptotically consistent, improve with increasing local neighborhood sizes, and have a high-dimensional error rate comparable to that of the centralized maximum likelihood estimator. (2) We present a family of latent variable Gaussian graphicalmodels whose marginal precision matrix has a “low-rank plus sparse” structure. Under mild conditions, we analyze the high-dimensional parameter error bounds for learning this family of models using regularized maximum likelihood estimation. (3) We consider a hypothesis testing framework for detecting emerging topics in topic models, and propose a novel surrogate test statistic for the standard likelihood ratio. By leveraging the theory of empirical processes, we prove asymptotic consistency for the proposed test and provide guarantees of the de
This paper proposes the use of estimation of distribution algorithms to deal with the problem of finding an optimal product of braid generators in topological quantum computing. We investigate how the regularities of ...
详细信息
ISBN:
(纸本)9783319135632;9783319135625
This paper proposes the use of estimation of distribution algorithms to deal with the problem of finding an optimal product of braid generators in topological quantum computing. We investigate how the regularities of the braid optimization problem can be translated into statistical regularities by means of the Boltzmann distribution. The introduced algorithm obtains solutions with an accuracy in the order of 10(-6), and lengths up to 9 times shorter than those expected from braids of the same accuracy obtained with other methods.
We develop a probabilisticgraphical model based methodology to efficiently perform uncertainty quantification in the presence of both stochastic input and multiple scales. Both the stochastic input and model response...
详细信息
We develop a probabilisticgraphical model based methodology to efficiently perform uncertainty quantification in the presence of both stochastic input and multiple scales. Both the stochastic input and model responses are treated as random variables in this framework. Their relationships are modeled by graphicalmodels which give explicit factorization of a high-dimensional joint probability distribution. The hyperparameters in the probabilistic model are learned using sequential Monte Carlo (SMC) method, which is superior to standard Markov chain Monte Carlo (MCMC) methods for multi-modal distributions. Finally, we make predictions from the probabilisticgraphical model using the belief propagation algorithm. Numerical examples are presented to show the accuracy and efficiency of the predictive capability of the developed graphical model. (C) 2013 Elsevier Inc. All rights reserved.
Gated Bayesian networks (GBNs) are a recently introduced extension of Bayesian networks that aims to model dynamical systems consisting of several distinct phases. In this paper, we present an algorithm for semi-autom...
详细信息
ISBN:
(纸本)9783319114330;9783319114323
Gated Bayesian networks (GBNs) are a recently introduced extension of Bayesian networks that aims to model dynamical systems consisting of several distinct phases. In this paper, we present an algorithm for semi-automatic learning of GBNs. We use the algorithm to learn GBNs that output buy and sell decisions for use in algorithmic trading systems. We show how using the learnt GBNs can substantially lower risks towards invested capital, while at the same time generating similar or better rewards, compared to the benchmark investment strategy buy-and-hold.
The Libra Toolkit is a collection of algorithms for learning and inference with discrete probabilisticmodels, including Bayesian networks, Markov networks, dependency networks, and sum-product networks. Compared to o...
详细信息
The Libra Toolkit is a collection of algorithms for learning and inference with discrete probabilisticmodels, including Bayesian networks, Markov networks, dependency networks, and sum-product networks. Compared to other toolkits, Libra places a greater emphasis on learning the structure of tractable models in which exact inference is efficient. It also includes a variety of algorithms for learning graphicalmodels in which inference is potentially intractable, and for performing exact and approximate inference. Libra is released under a 2-clause BSD license to encourage broad use in academia and industry.
Many human diseases including cancer are the result of perturbations to transcriptional regulatory networks that control context-specific expression of genes. A comparative approach across multiple cancer types is a p...
详细信息
Many human diseases including cancer are the result of perturbations to transcriptional regulatory networks that control context-specific expression of genes. A comparative approach across multiple cancer types is a powerful approach to illuminate the common and specific network features of this family of diseases. Recent efforts from The Cancer Genome Atlas (TCGA) have generated large collections of functional genomic data sets for multiple types of cancers. An emerging challenge is to devise computational approaches that systematically compare these genomic data sets across different cancer types that identify common and cancer-specific network components. We present a module-and network-based characterization of transcriptional patterns in six different cancers being studied in TCGA: breast, colon, rectal, kidney, ovarian, and endometrial. Our approach uses a recently developed regulatory network reconstruction algorithm, modular regulatory network learning with per gene information (MERLIN), within a stability selection framework to predict regulators for individual genes and gene modules. Our module-based analysis identifies a common theme of immune system processes in each cancer study, with modules statistically enriched for immune response processes as well as targets of key immune response regulators from the interferon regulatory factor (IRF) and signal transducer and activator of transcription (STAT) families. Comparison of the inferred regulatory networks from each cancer type identified a core regulatory network that included genes involved in chromatin remodeling, cell cycle, and immune response. Regulatory network hubs included genes with known roles in specific cancer types as well as genes with potentially novel roles in different cancer types. Overall, our integrated module and network analysis recapitulated known themes in cancer biology and additionally revealed novel regulatory hubs that suggest a complex interplay of immune response, cell cycle
暂无评论