We derive a PAC-Bayesian generalization bound for density estimation. Similar to the PAC-Bayesian generalization bound for classification, the result has the appealingly simple form of a tradeoff between empirical performance and the KL-divergence of the posterior from the prior. Moreover, the PACBayesian generalization bound for classification can be derived as a special case of the bound for ...

Information theoretic measures are of paramount importance in Bayesian inference. Yet, their direct application in practice has been somewhat limited by the severe computational diiculties encountered in complex problems. In this paper we discuss implementation strategies for fast numerical computations of Entropies and Kullback-Leibler divergences that are relevant to Bayesian inference and de...

Multiple elements such as lighting, colors, dialogue, and camera motion contribute to the style of a movie. Among them, camera motion is commonly overlooked yet a crucial point. For instance, documentaries tend to use long smooth pans whereas action movies usually have short and dynamic movements. This information, also referred to as global motion, could be leveraged by various applications in...

This paper offers a methodological contribution for computing the distance between two empirical distributions in an Euclidean space of very large dimension. We propose to use decision trees instead of relying on standard quantifi10 cation of the feature space. Our contribution is two-fold: We first define a new distance between empirical distributions, based on the Kullback-Leibler (KL) diverg...

A technique is described to prove termination of constructor systems (CSs) automatically. The technique consists of three major steps. First, determine the dependency pairs of a constructor system. Second, nd an equational theory in which the constructor system is contained, and third, prove that no in nite chain w.r.t. the equational theory of these dependency pairs exists. The rst step is eas...

In this work, speech parameterization based on the continuous multiresolution divergence is used to modify a textindependent phone segmentation algorithm. This encoding is employed as input and also replaces an stage of the segmentation procedure responsible for the estimation of the intensity of changes in signal features. The segmentation performance of this representation has been compared w...

Entropy is used as a quantitative measurement of the information deficiency. In uncertain set theory, logarithm entropy for uncertain set has been proposed. However, it fails to measure the degree of uncertainty associated with some uncertain sets. Thus this paper will propose sine entropy for uncertain set as a supplement, and investigate its properties such as translation invariance and posit...

We study the extraction of characteristics of user behavior in video session encoded as stochastic matrices of finite Markov chain. These behaviors are clustered using a dissimilarity based on the Kullbach-Leibler divergence between probability distributions. The center of each cluster is regarded as the model that generates the behaviors assigned to the cluster. This choice is based on the rel...

A technique to prove termination of term rewrite systems, or more precise, constructorsystems (CSs) is presented.Soundnessof this technique is proved and it is described how the technique can be performed automatically for a subclass of the CSs. Whereas simplification orders fail in proving termination of CSs that are not simply terminating, the presented technique may still automatically prove...