Bayesian networks have received much attention in the recent literature. In this article, we propose an approach to learn Bayesian networks using the stochastic approximation Monte Carlo (SAMC) algorithm. Our approach...
详细信息
Bayesian networks have received much attention in the recent literature. In this article, we propose an approach to learn Bayesian networks using the stochastic approximation Monte Carlo (SAMC) algorithm. Our approach has two nice features. Firstly, it possesses the self-adjusting mechanism and thus avoids essentially the local-trap problem suffered by conventional MCMC simulation-based approaches in learning Bayesian networks. Secondly, it falls into the class of dynamic importance sampling algorithms;the network features can be inferred by dynamically weighted averaging the samples generated in the learning process, and the resulting estimates can have much lower variation than the single model-based estimates. The numerical results indicate that our approach can mix much faster over the space of Bayesian networks than the conventional MCMC simulation-based approaches. (C) 2008 Elsevier B.V. All rights reserved.
Label-free relative quantitative proteornics is a powerful tool for the survey of protein level changes between two biological samples. We have developed and applied an algorithm using chromatographic alignment of mu ...
详细信息
Label-free relative quantitative proteornics is a powerful tool for the survey of protein level changes between two biological samples. We have developed and applied an algorithm using chromatographic alignment of mu LC-MS runs to improve the detection of differences between complex protein mixtures. We demonstrate the performance of our software by finding differences in E. coli protein abundance upon induction of the lac operon genes using isopropyl P-D-thiogalactopyranoside. The use of our alignment gave a 4-fold decrease in mean relative retention time error and a 6-fold increase in the number of statistically significant differences between samples. Using a conservative threshold, we have identified 5290 total mu LC-MS regions that have a different abundance between these samples. Of the detected difference regions, only 23% were mapped to MS/MS peptide identifications. We detected 74 proteins that had a greater relative abundance in the induced sample and 21 with a greater abundance in the uninduced sample. We have developed an effective tool for the label-free detection of differences between samples and demonstrate an increased sensitivity following chromatographic alignment.
Consider a distributed system N in which each agent has an input value and each communication link has a weight. Given a global function, that is, a function f whose value depends on the whole network, the goal is for...
详细信息
Consider a distributed system N in which each agent has an input value and each communication link has a weight. Given a global function, that is, a function f whose value depends on the whole network, the goal is for every agent to eventually compute the value f (N). We call this problem global function computation. Various solutions for instances of this problem, such as Boolean function computation, leader election, (minimum) spanning tree construction, and network determination, have been proposed, each under particular assumptions about what processors know about the system and how this knowledge can be acquired. We give a necessary and sufficient condition for the problem to be solvable that generalizes a number of well-known results (Attyia et al. in J ACM 35(4):845-875, 1988;Yamashita and Kameda in IEEE Trans Parallel Distrib Syst 7(1):69-89, 1996;Yamashita and Kameda in IEEE Trans Parallel Distrib Syst 10(9):878-887, 1999). We then provide a knowledge-based (kb) program (like those of Fagin et al. (Reasoning about knowledge, MIT Press, Cambridge, 1995, Distrib Comput 10(4):199-225, 1997)) that solves global function computation whenever possible. Finally, we improve the message overhead inherent in our initial kb program by giving a counterfactual belief-based program (Halpern and Moses in Distrib Comput 17(2):91-106, 2004) that also solves the global function computation whenever possible, but where agents send messages only when they believe it is necessary to do so. The latter program is shown to be implemented by a number of well-known algorithms for solving leader election.
The constitutive equation of an Euler-Bernoulli beam under the excitation of moving mass is considered. The dynamics of the uncontrolled system is governed by a linear, self-adjoint partial differential equation. A Di...
详细信息
The constitutive equation of an Euler-Bernoulli beam under the excitation of moving mass is considered. The dynamics of the uncontrolled system is governed by a linear, self-adjoint partial differential equation. A Dirac-delta function is used to describe the position of the moving mass along the beam and its inertial effects. An approximate formulation to the problem is obtained by limiting the inertial effect of the moving mass merely to the vertical component of acceleration. Having defined a "critical velocity" in terms of the fundamental period and span of the beam, it is shown that for smaller velocities, the approximate and exact approaches to the problem almost coincide. Since, the defined critical velocity is fairly large compared to those in practical cases, the approximate approach can effectively be used for a wide range of problems. There is however a slight variation in critical velocity depending on the weight of the moving mass. On the other hand, it is shown that the effect of higher vibrational modes is not negligible for certain velocity ranges. Finally, a linear classical optimal control algorithm with a time varying gain matrix with displacement-velocity feedback is used to control the response of the beam. The efficiency of the control algorithm in suppressing the response of the system under the effect of moving mass with different number of controlled modes and actuators is investigated. (c) 2007 Elsevier Ltd. All rights reserved.
Cartograms are used to visualize geographically distributed data by scaling the regions of a map (e.g., US states) such that their areas are proportional to some data associated with them (e.g., population). Thus the ...
详细信息
Cartograms are used to visualize geographically distributed data by scaling the regions of a map (e.g., US states) such that their areas are proportional to some data associated with them (e.g., population). Thus the cartogram computation problem can be considered as a map deformation problem where the input is a planar polygonal map M and an assignment of some positive weight for each region. The goal is to create a deformed map M, where the area of each region realizes the weight assigned to it (no cartographic error) while the overall map remains readable and recognizable (e.g., the topology, relative positions and shapes of the regions remain as close to those before the deformation as possible). Although several such measures of cartogram quality are well-known, different cartogram generation methods optimize different features and there is no standard set of quantitative metrics. In this paper we define such a set of seven quantitative measures, designed to evaluate how faithfully a cartogram represents the desired weights and to estimate the readability of the final representation. We then study several cartogram-generation algorithms and compare them in terms of these quantitative measures.
The logic of equalities with Uninterpreted Functions is used in the formal verification community mainly for proofs of equivalence: proving that two versions of a hardware design are the same, or that input and output...
详细信息
The logic of equalities with Uninterpreted Functions is used in the formal verification community mainly for proofs of equivalence: proving that two versions of a hardware design are the same, or that input and output of a compiler are semantically equivalent are two prominent examples of such proofs. We introduce a new decision procedure for this logic that generalizes two leading decision procedures that were published in the last few years: the Positive Equality approach suggested by Bryant et al. [Exploiting positive equality in a logic of equality with uninterpreted functions, in: Proc. 11th Intl. Conference on computer Aided Verification (CAV'99), 1999], and the Range-Allocation algorithm suggested by Pnueli et al. [The small model property: how small can it be? Information and Computation 178 (1) (2002) 279-293]. Both of these methods reduce this logic to pure Equality Logic (without Uninterpreted Functions), and then, due to the small model property that such formulas have, find a small domain to each variable that is sufficiently large to maintain the satisfiability of the formula. The state-space spanned by these domains is then traversed with a BDD-based engine. The Positive Equality approach identifies terms that have a certain characteristic in the original formula (before the reduction to pure Equality Logic), and replaces them with unique constants. The Range-Allocation algorithm analyzes the structure of the formula after the reduction to equality logic with a graph-based procedure to allocate a small set of values to each variable. The former, therefore, has an advantage when a large subset of the terms can be replaced with constants, and disadvantage in the other cases. In this paper we essentially merge the two methods, while improving both with a more careful analysis of the formula's structure. We show that the new method is provably dominant over both methods, theoretically as well as empirically. (c) 2005 Elsevier Inc. All rights reserved.
A novel strategy for model predictive control tuning considering nonsquare systems, with more outputs than inputs, and set points or ranges for controlled variables, is introduced. This methodology is applicable for a...
详细信息
A novel strategy for model predictive control tuning considering nonsquare systems, with more outputs than inputs, and set points or ranges for controlled variables, is introduced. This methodology is applicable for any predictive control algorithm, since it has adjustable parameters according to the implementation, and is divided in three parts that consists of, initially, the determination of the best attainable closed-loop performance function for the multiscenarios, generated from the complete model, including performance and robustness metrics as additional constraints. Based on the attainable closed-loop function, the second step is the calculation of the optimal scaling for the process and, finally, the controller weights are tuned leading the process to the best operating condition. This constrained strategy was applied in the controller design of the Shell heavy oil fractionator benchmark showing good results for zone region tracking.
Based on the optimal fusion algorithm weighted by matrices in the linear minimum variance (LMV) sense, a distributed optimal information fusion for the steady-state Kalman multi-step predictor is given for discrete li...
详细信息
Based on the optimal fusion algorithm weighted by matrices in the linear minimum variance (LMV) sense, a distributed optimal information fusion for the steady-state Kalman multi-step predictor is given for discrete linear stochastic control systems with multiple sensors and correlated noises, where the same sample period is assumed. When the noise statistics information is unknown, the distributed information fusion estimators for the noise statistics parameters are presented based on the correlation functions and the weighting average approach. Further, a self-tuning information fusion multi-step predictor is obtained. It has a two-stage fusion structure. The first-stage fusion is to obtain the fused noise statistics information. The second-stage fusion is to obtain the fused multi-step predictor. A simulation example shows the effectiveness.
A description of a real algebraic variety in a"e(3) is given. This variety plays an important role in the investigation of the Einstein metrics whose evolution is studied using the normalized Ricci flow. To revea...
详细信息
A description of a real algebraic variety in a"e(3) is given. This variety plays an important role in the investigation of the Einstein metrics whose evolution is studied using the normalized Ricci flow. To reveal the internal structure of this variety, a description of all its singular points is given. Due to the internal symmetry of this variety, a part of the investigation uses elementary symmetric polynomials. All the computations are performed using computer algebra algorithms (in particular, Grobner bases) and algorithms for dealing with polynomial ideals. As an auxiliary result, a proposition about the structure of the discriminant surface of a cubic polynomial is proved.
This paper proposes a human tracking approach in a distributed wireless sensor network. Most of the efforts on human tracking focus on vision techniques. However, most vision-based approaches to moving object detectio...
详细信息
This paper proposes a human tracking approach in a distributed wireless sensor network. Most of the efforts on human tracking focus on vision techniques. However, most vision-based approaches to moving object detection involve intensive real-time computations. In this paper, we present an algorithm for human tracking using low-cost range wireless sensor nodes which can contribute lower computational burden based on a distributed computing system, while the centralized computing system often makes some information from sensors delay. Because the human target often moves with high maneuvering, the proposed algorithm applies the interacting multiple model (IMM) filter techniques and a novel sensor node selection scheme developed considering both the tracking accuracy and the energy cost which is based on the tacking results of IMM filter at each time step. This paper also proposed a novel sensor management scheme which can manage the sensor node effectively during the sensor node selection and the tracking process. Simulations results show that the proposed approach can achieve superior tracking accuracy compared to the most recent human motion tracking scheme.
暂无评论