Results 1  10
of
15
PrivacyPreserving Multivariate Statistical Analysis: Linear Regression and Classification
 In Proceedings of the 4th SIAM International Conference on Data Mining
, 2004
"... analysis technique that has found applications in various areas. In this paper, we study some multivariate statistical analysis methods in Secure 2party Computation (S2C) framework illustrated by the following scenario: two parties, each having a secret data set, want to conduct the statistical ana ..."
Abstract

Cited by 60 (1 self)
 Add to MetaCart
analysis technique that has found applications in various areas. In this paper, we study some multivariate statistical analysis methods in Secure 2party Computation (S2C) framework illustrated by the following scenario: two parties, each having a secret data set, want to conduct the statistical analysis on their joint data, but neither party is willing to disclose its private data to the other party or any third party. The current statistical analysis techniques cannot be used directly to support this kind of computation because they require all parties to send the necessary data to a central place. In this paper, We define two Secure 2party multivariate statistical analysis problems: Secure 2party Multivariate Linear Regression problem and Secure 2party Multivariate Classification problem. We have developed a practical security model, based on which we have developed a number of building blocks for solving these two problems.
Interpolation, Spectrum Analysis, ErrorControl Coding, and FaultTolerant Computing
 In Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 97, volume III
, 1997
"... This paper uncovers relations between the topics mentioned in the title, relations that we believe to have gone nearly unnoticed so far. More precisely, we show that four often studied problems in signal processing, spectrum analysis, information theory, and computing are closely related or even equ ..."
Abstract

Cited by 16 (7 self)
 Add to MetaCart
This paper uncovers relations between the topics mentioned in the title, relations that we believe to have gone nearly unnoticed so far. More precisely, we show that four often studied problems in signal processing, spectrum analysis, information theory, and computing are closely related or even equivalent in a certain sense (if one of them can be solved, so can any of the others, and using essentially the same algorithms). The problems are (i) a nonlinear bandlimited finitedimensional interpolation problem (ii) the problem of estimating a signal that is the superposition of a finite number of harmonics (iii) an errorcontrol coding problem in the real field, and (iv) certain techniques that occur in algorithmbased fault tolerant computing. The advantages of recognizing these problems as equivalent are obvious: the techniques commonly used in one field can be imported to the others, the duplication of research e#orts is prevented, and the overall degree of understanding of the four problems increases. New algorithms are suggested as a result of these investigations. 1. NOTATION The complex ndimensional space, with the usual inner product and norm, is denoted by C n . A signal is a ndimensional complex vector x, with components, or samples, x(0), x(1), . . . , x(n 1). The Fourier matrix F is the n n matrix whose elements F ab are given by F ab = e j n ab where j denotes the imaginary unit. The discrete Fourier transform (DFT) of x, denoted by x, is defined by x = Fx. A signal x is bandlimited if a subset of the samples of x vanish, and is lowpass if the nonzero DFT Fax +35134370545, emails vieira@inesca.pt and pjf@inesca.pt. This work was supported by JNICT.
NonConcurrent Error Detection and Correction in FaultTolerant Linear FiniteState Machines
 IEEE Transactions on Automatic Control
, 2002
"... Previous work constructed faulttolerant linear finitestate machines (LFSMs) by embedding a given LFSM into a larger, redundant LFSM that preserves the evolution and properties of the original one while enabling an external mechanism to perform concurrent error detection and correction. In this pap ..."
Abstract

Cited by 13 (5 self)
 Add to MetaCart
Previous work constructed faulttolerant linear finitestate machines (LFSMs) by embedding a given LFSM into a larger, redundant LFSM that preserves the evolution and properties of the original one while enabling an external mechanism to perform concurrent error detection and correction. In this paper, we construct faulttolerant LFSMs that allow the external mechanism to perform nonconcurrent error detection and correction (i.e., to perform checking periodically, for instance, once every N time steps). This approach relaxes the requirements on the reliability of the error detecting/correcting mechanism because the mechanism can operate at a slower speed than the rest of the system. We characterize constructions for nonconcurrent detection of single errors and also present schemes which use BCH coding to allow for efficient nonconcurrent detection and correction of multiple errors.
Mathematics for Multimedia Signal Processing II: Discrete Finite Frames and Signal Reconstruction
 in Signal Processing for Multimedia
, 1999
"... . Certain signal reconstruction problems can be understood in terms of frames and redundant representations. The redundancy is useful because it leads to robust signal representations, that is, representations in which partial loss of data can be tolerated without misbehavior or adverse effects. ..."
Abstract

Cited by 10 (3 self)
 Add to MetaCart
. Certain signal reconstruction problems can be understood in terms of frames and redundant representations. The redundancy is useful because it leads to robust signal representations, that is, representations in which partial loss of data can be tolerated without misbehavior or adverse effects. This chapter begins by presenting a few engineering problems in which robust data representations play a central role. It turns out that these problems, which occur in signal processing, spectrum analysis, information theory, and faulttolerant computing, are closely related or even equivalent. However, perhaps surprisingly, the connections between them have gone nearly unnoticed so far. Frames, and in particular discrete finite frames, provide one of the ways of understanding certain of these problems, including the important missing data problem. Some of the methods that can be used to recover from missing data errors are examined, emphasizing finitedimensional theory because of i...
Structured Redundancy for Fault Tolerance in LTI StateSpace Models and Petri Nets
 Kybernetika
, 1999
"... The design and implementation of dynamic systems has traditionally focused on minimal representations which require the least number of state variables. However, \structured redundancy"  redundancy that has been intentionally introduced in some systematic way  can be extremely important when fault ..."
Abstract

Cited by 9 (9 self)
 Add to MetaCart
The design and implementation of dynamic systems has traditionally focused on minimal representations which require the least number of state variables. However, \structured redundancy"  redundancy that has been intentionally introduced in some systematic way  can be extremely important when fault tolerance is desired. The redundancy can be used to detect and correct errors or to guarantee desirable performance despite hardware or computational failures. Modular redundancy, the traditional approach to fault tolerance, is prohibitively expensive because of the overhead in replicating the hardware. This paper discusses alternative methods for systematically introducing redundancy in dynamic systems. Our approach consists of mapping the state space of the original system into a redundant space of higher dimension while preserving the properties of the original system in some encoded form within this larger space. We illustrate our approach by focusing on linear timeinvariant (LTI) dyna...
Numerically Stable Real Number Codes Based on Random Matrices
 In Proceeding of the 5th International Conference on Computational Science (ICCS2005
, 2004
"... Abstract. Error correction codes defined over realnumber field have been studied and recognized as useful in many applications. However, most realnumber codes in literature are quite suspect in their numerical stability. In this paper, we introduce a class of realnumber codes based on random gene ..."
Abstract

Cited by 5 (5 self)
 Add to MetaCart
Abstract. Error correction codes defined over realnumber field have been studied and recognized as useful in many applications. However, most realnumber codes in literature are quite suspect in their numerical stability. In this paper, we introduce a class of realnumber codes based on random generator matrices over realnumber fields. Codes over complexnumber field are also discussed. Experiment results demonstrate our codes are numerically much more stable than existing codes in literature. 1
Fault tolerant QRDecomposition Algorithm Based on Householder
 Reflections and its Parallel Implementation, Proc.4th Int. Workshop Parallel Numerics`97
, 1997
"... Abstract. A faulttolerant algorithms based on Givens rotations and modified weighted checksum methods are proposed for matrix QRdecomposition. The purpose is to detect and correct the calculation errors occurred due to transient hardware faults during computation. The proposed algorithm enables to ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
Abstract. A faulttolerant algorithms based on Givens rotations and modified weighted checksum methods are proposed for matrix QRdecomposition. The purpose is to detect and correct the calculation errors occurred due to transient hardware faults during computation. The proposed algorithm enables to correct a single error in each row or column of an input matrix A(M,N) occurred at any among N steps of algorithm implementation. Consequently, it is possible to correct up to N 2 single transient errors during solving the whole decomposition problem. This effect is obtained by increasing the computational complexity of the original Givens method on 2,5N 2 + O(N) multiplyadd operations. Finally, the parallel version of proposed algorithm is designed, dedicated for realisation on a fixedsize linear processor array with fully local communications and low I/O requirements. 1
Numerically stable realnumber codes based on random matrices
 in ITW2004
"... Abstract — Error correction codes defined over realnumber and complexnumber fields have been studied and recognized as useful in many applications. However, most realnumber and complexnumber codes in literature are quite suspect in their numerical stability. In this paper, we introduce a class of ..."
Abstract

Cited by 3 (2 self)
 Add to MetaCart
Abstract — Error correction codes defined over realnumber and complexnumber fields have been studied and recognized as useful in many applications. However, most realnumber and complexnumber codes in literature are quite suspect in their numerical stability. In this paper, we introduce a class of numerically stable realnumber and complexnumber codes that are based on random generator matrices over realnumber and complexnumber fields.
An Experimental Evaluation of the Effectiveness of Automatic RuleBased Transformations for SafetyCritical Applications
 DFT'00, IEEE International Symposium on Defect and Fault Tolerance in VLSI Systems, 2000
, 2000
"... 1 Over the last years, an increasing number of safetycritical tasks have been demanded to computer systems. In particular, safetycritical computerbased applications are hitting markets where costs is a major issue, and thus solutions are required which conjugate fault tolerance with low costs. I ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
1 Over the last years, an increasing number of safetycritical tasks have been demanded to computer systems. In particular, safetycritical computerbased applications are hitting markets where costs is a major issue, and thus solutions are required which conjugate fault tolerance with low costs. In this paper, a softwarebased approach for developing safetycritical applications is analyzed. By exploiting an adhoc tool implementing the proposed technique, several benchmark applications have been hardened against transient errors. Fault Injection campaigns have been performed to evaluate the fault detection capability of the hardened applications. Moreover, a comparison of the proposed techniques with the AlgorithmBased Fault Tolerance (ABFT) approach is proposed. Experimental results show that the proposed approach is far more effective than ABFT in terms of fault detection capability when injecting transient faults in data and code memory, at a cost of an increased memory overhead. Moreover, the performance penalty introduced by the proposed technique is comparable, and sometimes lower, than that ABFT requires. 1.
Probabilistic Compensation for Digital Filters under Pervasive NoiseInduced Operator Errors, accepted
 in VLSI Test Symposium
, 2007
"... It is well known that scaled CMOS technologies are increasingly susceptible to induced soft errors and environmental noise. Probabilistic checksumbased error detection and compensation has been proposed in the past for scaled DSP circuits for which a certain level of inaccuracy can be tolerated as ..."
Abstract

Cited by 2 (1 self)
 Add to MetaCart
It is well known that scaled CMOS technologies are increasingly susceptible to induced soft errors and environmental noise. Probabilistic checksumbased error detection and compensation has been proposed in the past for scaled DSP circuits for which a certain level of inaccuracy can be tolerated as long as systemlevel QualityofService (QoS) metrics are satisfied. Although the technique has been shown to be effective in improving the SNR of digital filters, it can only handle errors that occur in the system states. However, the transienterror rate of combinational logic is increasing with technology scaling. Therefore, handling errors in the arithmetic logic circuitry of DSP systems is also essential. This is a significantly more difficult task due to the fact that a single error at the output of an adder or multiplier can propagate to more than one system state causing multiple states to be erroneous. In this paper, a unified scheme that can address probabilistic compensation for errors both in the system states and in the embedded adders and multipliers of DSP filters is developed. It is shown that by careful checksum code design, significant SNR improvements (up to 13 dB) can be obtained for linear filters in the presence of soft errors. 1.