For large vocabulary recognition system, as well as for flexible vocabulary applications using hidden Markov model(HMM), parameter smoothing and tying have been used to increase the reliability of models. This paper describes bottom-up and topdown clustering techniques for state level tying. This paper also describes a method of applying parameter smoothing to the clustered states and covarianc...

We consider a class of smoothing methods for minimization problems where the feasible set is convex but the objective function is not convex, not differentiable and perhaps not even locally Lipschitz at the solutions. Such optimization problems arise from wide applications including image restoration, signal reconstruction, variable selection, optimal control, stochastic equilibrium and spheric...

MAP (maximum a posteriori) reconstructions of transmission images produce more accurate ACFs (attenuation correction factors) than smoothed division of blank and transmission scans and analytical methods. Also, for highly speciÞc tracers, transmission images can provide useful anatomical cues for use in localizing structures and image coregistration. The resolution of MAP images reconstructed w...

Nonparametric regression, often called smoothing, is a widely used data analysis method. The use of a smoother requires the choice of a smoothing parameter that by balancing Ž delity and roughness controls how much smoothing is done. Two popular selection criteria for choosing the smoothing parameter are Cp and generalized maximum likelihood (GML). Each of these has its own problems. For Cp , t...

We study evolution of various statistical quantities of smoothed cosmic density and velocity fields using N-body simulations. The parameter C ≡ 〈V 2δ〉 /(〈V 2〉 〈δ2〉) characterizes nonlinear coupling of these two fields and determines behavior of bulk velocity dispersion as a function of local density contrast. It is found that this parameter depends strongly on the smoothing scale even in quasi-...