In this paper we present a visual learning approach that uses non-parametric probability estimators. We use entropy analysis over the training set in order to select the features that best represent the pattern class of faces, and set up discrete probability models. These models are tested in the context of maximum likelihooddetection of faces. Excellent results are reported in terms of the cor...

We introduce an estimator for the population mean based on maximizing likelihoods formed by parameterizing a kernel density estimate. Due to these origins, we have dubbed the estimator the maximum kernel likelihood estimate (mkle). A speedy computational method to compute the mkle based on binning is implemented in a simulation study which shows that the mkle at an optimal bandwidth is decidedl...

We present the CEM (Conditional Expectation Maximization) algorithm as an extension of the EM (Expectation Maximization) algorithm to conditional density estimation under missing data. A bounding and maximization process is given to speciically optimize conditional likelihood instead of the usual joint likelihood. We apply the method to conditioned mixture models and use bounding techniques to ...

In this paper, we review an extension of the learning rules in a Principal Component Analysis network which has been derived to be optimal for a specific probability density function. We note that this probability density function is one of a family of pdfs and investigate the learning rules formed in order to be optimal for several members of this family. We show that, whereas previous authors...

Maximnm likelihood estimates are reported to be best under all circumstances. Yet there are numerous simple examples where they plainly misbehave. One gives some eranmples for problems that had not been invented for the purpose of annoying ms,aximunm likelihood fans. Another example, imitated from B'hadu'r, has been specially created with just such a purpose in mind. Next, we present a list of ...

One popular class of such models are Markov networks, which use an undirected graph to represent dependencies among variables. Markov networks of low tree-width (i.e. having a triangulation with small cliques ) allow efficient computations, and are useful as learned probability models [8]. A well studied case is that in which the dependency structure is known in advance. In this case the underl...

We put forward a new method of growing regression trees via maximum likelihood. It inherits the CART (Brieman et al., 1984) backward fitting idea. However, standard likelihood based methods such as model selection criteria and likelihood ratio tests are naturally incorporated into each stage of the tree procedure. Compared with other least squared tree methods, maximum likelihood regression tre...

We present a novel method to recontruct the mass distribution of galaxy clusters from their gravitational lens effect on background galaxies. The method is based on a least-χ fit of the two-dimensional gravitational cluster potential. The method combines information from shear and magnification by the cluster lens and is designed to easily incorporate possible additional information. We describ...

Looking myopically at the larger features of the likelihood function, absent some fine detail, can theoretically improve maximum likelihood estimation. Such estimators are, in fact, used routinely, since numerical techniques for maximizing a computationally expensive likelihood function or for maximizing a Monte Carlo approximation to a likelihood function may be unable to investigate small sca...