The problem of estimating the intensity process of a doubly stochastic Poisson process is analyzed. Using covariance information, a recursive linear minimum mean-square error estimate is designed. Moreover, an efficient procedure for the computation of its associated error covariance is shown. The proposed solution becomes an alternative approach to the Kalman filter which is applicable under t...

We address the problem of training the free parameters of a statistical machine translation system. We show significant improvements over a state-of-the-art minimum error rate training baseline on a large ChineseEnglish translation task. We present novel training criteria based on maximum likelihood estimation and expected loss computation. Additionally, we compare the maximum a-posteriori deci...

In this paper a duality-based a posteriori error analysis is developed for the conforming hp Galerkin finite element approximation of second-order elliptic problems. Duality arguments combined with Galerkin orthogonalty yield representations of the error in arbitrary quantities of interest. From these error estimates, criteria are derived for the simultaneous adaptation of the mesh size h and t...

Digital Object Identifier: 10.1109/ TNN.2009.2025946 “Convergence analysis of the online BP training algorithm was proved using two fundamental theorems: One theorem claims that under mild conditions, the gradient sequence of the error function will converge to zero (the weak convergence), and another theorem concludes the convergence of the weight sequence defined by the procedure to a fixed v...

The Blind Source Separation (BSS) problem is often solved by maximizing objective functions reflecting the statistical dependency between outputs. Since global maximization may be difficult without exhaustive search, criteria for which it can be proved that all the local maxima correspond to an acceptable solution of the BSS problem have been developed. These criteria are used in a deflation pr...

The objective of pool-based incremental active learning is to choose a sample to label from a pool of unlabeled samples in an incremental manner so that the generalization error is minimized. In this scenario, the generalization error often hits a minimum in the middle of the incremental active learning procedure and then it starts to increase. In this paper, we address the problem of early lab...

Resolution of binary mixtures of theophylline (THEO), montelukast (MKST) and loratadine (LORA) with minimum sample pretreatment and without analyte separation has been successfully achieved by multivariate spectrophotometric calibration, together with partial least-squares (PLS-1), principal component regression (PCR) and hybrid linear analysis (HLA). Data of analysis were obtained from UV–Vis ...