In this work automatic methods for determining the number of gaussians per state in a set of Hidden Markov Models are studied. Four different mix-up criteria are proposed to decide how to increase the size of the states. These criteria, derived from Maximum Likelihood scores, are focused to increase the discrimination between states obtaining different number of gaussians per state. We compare ...

As the number of cores in modern processors keeps growing, power management at the core level is becoming increasingly important. One common method used to trade off power against performance is changing the clock frequency of the processor [1]. Due to time dependent variations in load for a single core and across multiple cores, dynamically adjustable, per core clock generators are highly desi...

Determinantal point processes (DPPs) provide a probabilistic formalism for modeling repulsive distributions over subsets. Such priors encourage diversity between selected items through the introduction of a kernel matrix that determines which items are similar and therefore less likely to appear together. We investigate the usefulness of such priors in the context of spike-and-slab variable sel...

A simple algorithm is described which determines the satisfiability over the reals of a conjunction of linear inequalities, none of which contains more than two variables. In the worst case the algorithm requires time O(WH#~Q~~+~ log n), where n is the number of variables and m the number of inequalities. Several considerations suggest that the algorithm may be useful in practice: it is simple ...

One of the problems with the traditional nite domains approach to solving integer problems in a constraint logic programming context is that all variables require explicit bounds. If no explicit bounds are available then the nite domain solver can be very ineecient on certain classes of problem. We present an alternative approach to solving integer constraints based on a polynomial-time solver ...

We consider a basic model for two essential on-Iioe decisions that have to be taken in workload modeis . The first is the decision to either continue or abort the service of a job. The second concerns the decision to either accept or reject new jobs. We show that under certsin regularity conditions, there exist optimal threshold policies for these two types of decisions.

In a least squares adjustment when more than one variable in an equation of condition has error, the results will be affected by unnecessary asymptotic bias unless the algorithm is properly formulated. Similar difficulties can be expected with robust estimation techniques that are based on extending least squares to a noneuclidean metric. This paper presents an algorithm for robust estimation i...