Results 1  10
of
148
Constructing Internet Coordinate System Based on Delay Measurement
, 2003
"... In this paper, we consider the problem of how to represent the locations of Internet hosts in a Cartesian coordinate system to facilitate estimate of the network distance between two arbitrary Internet hosts. We envision an infrastructure that consists of beacon nodes and provides the service of est ..."
Abstract

Cited by 121 (3 self)
 Add to MetaCart
(Show Context)
In this paper, we consider the problem of how to represent the locations of Internet hosts in a Cartesian coordinate system to facilitate estimate of the network distance between two arbitrary Internet hosts. We envision an infrastructure that consists of beacon nodes and provides the service of estimating network distance between two hosts without direct delay measurement. We show that the principal component analysis (PCA) technique can e#ectively extract topological information from delay measurements between beacon hosts. Based on PCA, we devise a transformation method that projects the distance data space into a new coordinate system of (much) smaller dimensions. The transformation retains as much topological information as possible and yet enables end hosts to easily determine their locations in the coordinate system. The resulting new coordinate system is termed as the Internet Coordinate System (ICS). As compared to existing work (e.g., IDMaps [1] and GNP [2]), ICS incurs smaller computation overhead in calculating the coordinates of hosts and smaller measurement overhead (required for end hosts to measure their distances to beacon hosts). Finally, we show via experimentation with reallife data sets that ICS is robust and accurate, regardless of the number of beacon nodes (as long as it exceeds certain threshold) and the complexity of network topology.
kPlane Clustering
 Journal of Global Optimization
, 2000
"... A finite new algorithm is proposed for clustering m given points in ndimensional real space into k clusters by generating k planes that constitute a local solution to the nonconvex problem of minimizing the sum of squares of the 2norm distances between each point and a nearest plane. The key to th ..."
Abstract

Cited by 76 (3 self)
 Add to MetaCart
(Show Context)
A finite new algorithm is proposed for clustering m given points in ndimensional real space into k clusters by generating k planes that constitute a local solution to the nonconvex problem of minimizing the sum of squares of the 2norm distances between each point and a nearest plane. The key to the algorithm lies in a formulation that generates a plane in ndimensional space that minimizes the sum of the squares of the 2norm distances to each of m1 given points in the space. The plane is generated by an eigenvector corresponding to a smallest eigenvalue of an n \Theta n simple matrix derived from the m1 points. The algorithm was tested on the publicly available Wisconsin Breast Prognosis Cancer database to generate well separated patient survival curves. In contrast, the kmean algorithm did not generate such wellseparated survival curves. 1 Introduction There are many approaches to clustering such as statistical [2, 9, 6], machine learning [7, 8] and mathematical programming [15...
Multiresolution Histograms and their Use for Recognition
 IEEE Trans. on PAMI
, 2004
"... ..."
(Show Context)
Numerical Methods for Simultaneous Diagonalization
 SIAM J. Matrix Anal. Applicat
, 1993
"... We present a Jacobilike algorithm for simultaneous diagonalization of commuting pairs of complex normal matrices by unitary similarity transformations. The algorithm uses a sequence of similarity transformations by elementary complex rotations to drive the offdiagonal entries to zero. We show th ..."
Abstract

Cited by 56 (0 self)
 Add to MetaCart
We present a Jacobilike algorithm for simultaneous diagonalization of commuting pairs of complex normal matrices by unitary similarity transformations. The algorithm uses a sequence of similarity transformations by elementary complex rotations to drive the offdiagonal entries to zero. We show that its asymptotic convergence rate is quadratic and that it is numerically stable. It preserves the special structure of real matrices, quaternion matrices and real symmetric matrices.
The CMU Reconfigurable Modular Manipulator System
 TECHNICAL REPORT
, 1988
"... Modular manipulator designs have long been b een considered for use as research tools, and as the basis for easily
modified industrial manipulators. In these manipulators the links and joints are discrete and modular components
that can be assembled into a desired manipulator configuration. As ha ..."
Abstract

Cited by 48 (15 self)
 Add to MetaCart
Modular manipulator designs have long been b een considered for use as research tools, and as the basis for easily
modified industrial manipulators. In these manipulators the links and joints are discrete and modular components
that can be assembled into a desired manipulator configuration. As hardware advances have made actual modular
manipulators practical, various capabilities of such manipulators have gained interest. Particularly desirable is the
ability to rapidly reconfigure such a manipulator, in order to custom tailor i t to specific tasks. T i reconfiguration
hs
greatly enhances the capability of a given amount of manipulator hardware. This paper discusses the development of
a prototype modular manipulator and the implementation of a configuration independent manipulator kinematics
algorithm used for path planning in the prototype.
A reliable and computationally efficient algorithm for imposing the saddle point property in dynamic models. Manuscript, Federal Reserve Board of Governors
"... linear saddle point models. The algorithm has proved useful in a wide array of applications including analyzing linear perfect foresight models, providing initial solutions and asymptotic constraints for nonlinear models. The algorithm solves linear problems with dozens of lags and leads and hundred ..."
Abstract

Cited by 39 (2 self)
 Add to MetaCart
linear saddle point models. The algorithm has proved useful in a wide array of applications including analyzing linear perfect foresight models, providing initial solutions and asymptotic constraints for nonlinear models. The algorithm solves linear problems with dozens of lags and leads and hundreds of equations in seconds. The technique works well for both symbolic algebra and numerical computation. Although widely used at the Federal Reserve, few outside the central bank know about or have used the algorithm. This paper attempts to present the current algorithm in a more accessible format in the hope that economists outside the Federal Reserve may also nd it useful. In addition, over the years there have been undocumented changes in approach that have improved the eciency and reliability of algorithm. This paper describes the present state of development of this set of tools.
Note On The Location Of Optimal Classifiers In NDimensional ROC Space
, 1999
"... The comparative study of classifier performance is a worthwhile concern in Machine Learning. Empirical comparisons typically examine unbiased estimates of predictive accuracy of different algorithms  the assumption being that the classifier with the highest accuracy would be the "optimal" ..."
Abstract

Cited by 38 (1 self)
 Add to MetaCart
(Show Context)
The comparative study of classifier performance is a worthwhile concern in Machine Learning. Empirical comparisons typically examine unbiased estimates of predictive accuracy of different algorithms  the assumption being that the classifier with the highest accuracy would be the "optimal" choice of classifier for the problem. The qualification on optimality is needed here, as choice is restricted to the classifiers being compared, and the estimates are typically subject to sampling errors. Comparisons based on predictive accuracy overlook two important practical concerns, namely (a) class distributions cannot be specified precisely. Distribution of classes in the training set are thus rarely matched exactly on new data; and (b) that the costs of different types of errors may be unequal. Using techniques developed in signal detection, Provost and Fawcett describe an elegant method for the comparative assessment of binary classifiers that takes these considerations into account. Thei...
A sensing chair using pressure distribution sensors
 IEEE/ASME TRANSACTIONS ON MECHATRONICS
, 2001
"... One challenge in multimodal interface research is the lack of robust subsystems that support multimodal interactions. By focusing on a chair—an object that is involved in virtually all human–computer interactions, the sensing chair project enables an ordinary office chair to become aware of its occ ..."
Abstract

Cited by 32 (1 self)
 Add to MetaCart
(Show Context)
One challenge in multimodal interface research is the lack of robust subsystems that support multimodal interactions. By focusing on a chair—an object that is involved in virtually all human–computer interactions, the sensing chair project enables an ordinary office chair to become aware of its occupant’s actions and needs. Surfacemounted pressure distribution sensors are placed over the seatpan and backrest of the chair for real time capturing of contact information between the chair and its occupant. Given the similarity between a pressure distribution map and a grayscale image, pattern recognition techniques commonly used in computer and robot vision, such as principal components analysis, have been successfully applied to solving the problem of sitting posture classification. The current static posture classification system operates in real time with an overall classification accuracy of 96 % and 79 % for familiar (people it had felt before) and unfamiliar users, respectively. Future work is aimed at a dynamic posture tracking system that continuously tracks not only steadystate (static) but transitional (dynamic) sitting postures. Results reported here form important stepping stones toward an intelligent chair that can find applications in many areas including multimodal interfaces, intelligent environment, and safety of automobile operations.
Multiple Description Decoding of Overcomplete Expansions Using Projections onto Convex Sets
 In Proceedings of Data Compression Conference
, 1999
"... This paper presents a POCSbased algorithm for consistent reconstruction of a signal x 2 R K from any subset of quantized coefficients y 2 R N in an N \Theta K overcomplete frame expansion y = Fx, N = 2K. By choosing the frame operator F to be the concatenation of two K \Theta K invertible tr ..."
Abstract

Cited by 28 (0 self)
 Add to MetaCart
(Show Context)
This paper presents a POCSbased algorithm for consistent reconstruction of a signal x 2 R K from any subset of quantized coefficients y 2 R N in an N \Theta K overcomplete frame expansion y = Fx, N = 2K. By choosing the frame operator F to be the concatenation of two K \Theta K invertible transforms, the projections may be computed in R K using only the transforms and their inverses, rather than in the larger space R N using the pseudoinverse as proposed in earlier work. This enables practical reconstructions from overcomplete frame expansions based on wavelet, subband, or lapped transforms of an entire image, which has heretofore not been possible. 1 Introduction Multiple description (MD) source coding is the problem of encoding a single source fX i g into N separate binary descriptions at rates R 1 ; : : : ; RN bits per symbol such that any subset S of the descriptions may be received and together decoded to an expected distortion D S commensurate with the total b...
Model reduction using proper orthogonal decomposition, 2011. Lecture Notes, University of Konstanz, www.math.unikonstanz.de/numerik/ personen/volkwein/teaching/PODVorlesung.pdf. Reduction for Parametrized PDEs 27 Andrea Manzoni CMCS – Modelling and Scie
 CMCS – Modelling and Scientific Computing MATHICSE – Mathematics Institute of Computational Science and Engineering EPFL – Ecole Polytechnique Fédérale de Lausanne Station 8, CH1015 Lausanne Switzerland and MOX – Modellistica e Calcolo Scientifico Dipart
, 2012
"... Abstract. In this lecture notes an introduction to model reduction utilizing proper orthogonal decomposition (POD) is given. The close connection between POD and singular value decomposition (SVD) of rectangular matrices is emphasized. As an application POD is used to derive a reducedorder model f ..."
Abstract

Cited by 25 (3 self)
 Add to MetaCart
(Show Context)
Abstract. In this lecture notes an introduction to model reduction utilizing proper orthogonal decomposition (POD) is given. The close connection between POD and singular value decomposition (SVD) of rectangular matrices is emphasized. As an application POD is used to derive a reducedorder model for nonlinear initial value problems. The strategy is extended to linearquadratic optimal control problems governed by ordinary differential equations. The relationship to classical model reduction techniques like balanced truncation is studied. 1. The POD method in Rm In this section we introduce the POD method in the Euclidean space Rm and study the close connection to the SVD of rectangular matrices; see [6]. We also refer to the monograph [3]. 1.1. POD and SVD. Let Y = [y1,..., yn] be a realvalued m × n matrix of rank d ≤ min{m,n} with columns yj ∈ Rm, 1 ≤ j ≤ n. Consequently, (1.1) y ̄ = 1 n n∑ j=1 yj can be viewed as the columnaveraged mean of the matrix Y. SVD [10] guarantees the existence of real numbers σ1 ≥ σ2 ≥... ≥ σd> 0 and orthogonal matrices U ∈ Rm×m with columns {ui}mi=1 and V ∈ Rn×n with columns {vi}ni=1 such that