We develop an adapted version of the Bayesian Personalized Ranking (BPR) optimization criterion (Rendle et al., 2009) that takes the non-uniform sampling of negative test items — as in track 2 of the KDD Cup 2011 — into account. Furthermore, we present a modified version of the generic BPR learning algorithm that maximizes the new criterion. We use it to train ranking matrix factorization model...

A new digital control method for a time varying sampled data system called a variable period deadbeat control is proposed to control high order nonholonomic chained systems with drift terms. This control law has a potential to avoid an excessive overshoot to prevent the trajectories close to the singular manifold of the transformation needed to obtain the chained forms. The control will be appl...

There are numerous methods for interpolating uniformly sampled, aliased seismic data, but few can handle the combination of nonuniform sampling and aliasing. We combine the principles of Fourier reconstruction of nonaliased, nonuniformly sampled data with the ideas of frequency-wavenumber f-k interpolation of aliased, uniformly sampled data in a new two-step algorithm. In the first step, we est...

Using Bayesian probability theory we demonstrate that the Lomb-Scargle periodogram may be generalized in a straightforward manner to nonuniformly nonsimultaneously sampled quadrature data when the sinusoid has arbitrary amplitude modulation. This generalized Lomb-Scargle periodogram is the sufficient statistic for single frequency estimation in a wide class of problems ranging from stationary f...

Given a parameterized (by sampling period T ) family of approximate discrete-time models of a sampled nonlinear plant and given a family of controllers stabilizing the family of plant models for all T sufficiently small, we present conditions which guarantee that the same family of controllers semi-globally-practically stabilizes the exact discrete-time model of the plant for sufficiently small...