A Maximum Entropy Approach for Collaborative Filtering

Collaborative filtering (CF) involves predicting the preferences of a user for a set of items given partial knowledge of the user's preferences for other items, while leveraging a database of profiles for other users. CF has applications e.g. in predicting Web sites a person will visit and in recommending products. Fundamentally, CF is a pattern recognition task, but a formidable one, often involving a huge feature space, a large data set, and many missing features. Even more daunting is the fact that a CF inference engine must be capable of predicting any (user-selected) items, given any available set of partial knowledge on the user's other preferences. In other words, the model must be designed to solve any of a huge (combinatoric) set of possible inference tasks. CF techniques include memory-based, classification-based, and statistical modelling approaches. Among these, modelling approaches scale best with large data sets and are the most adept at handling missing features. The disadvantage of these methods lies in the statistical assumptions (e.g. feature independence), which may be unjustified. To address this shortcoming we propose a new model-based CF method, based on the maximum entropy principle. For the MS Web application, the new method is demonstrated to outperform a number of CF approaches, including naive Bayes and latent variable (cluster) models, support vector machines (SVMs), and the (Pearson) correlation method.     

Extended fundamental frequency analysis of the LCC resonant converter

Fundamental frequency techniques are used to analyze the series-parallel resonant converter under heavy load conditions, both with a continuous, but distorted parallel capacitor voltage waveform, and with a discontinuous capacitor voltage waveform. The analysis is validated with results from an experimental prototype. The application of the technique to the parallel-loaded L-C resonant converter is also considered.     

Numerical simulation of ion-exchange in glass for integratedoptical components

A numerical model based on the finite-element method for solving the nonlinear diffusion equation of ion-exchange in glasses is presented. The model is applied to the calculation of the concentration profile for a Ag⁺-Na⁺ ion-exchanged channel waveguide. The calculated concentration profile is in good agreement with a measured refractive-index profile of a waveguide fabricated with the same parameters as used in the calculations     

Quadratic solitons: recent developments

Multicolor optical solitons mediated by quadratic nonlinearities have been observed experimentally during the last few years in a variety of materials and geometries in a number of laboratories around the world. Today, many of the basic properties of the corresponding soliton families are theoretically well established. Nevertheless, experiments continuously reveal new surprises, and the topic is entering new theoretical and experimental territories. In this paper, we briefly outline the key features of the quadratic solitons and overview the latest developments.     

Degradation of junction leakage in devices subjected to gateoxidation in nitrous oxide [MOS transistors]

Nitrided gate oxides offer several electrical and reliability advantages over conventional oxides and also provide a good barrier against impurity diffusion. Oxidation in nitrous oxide (N₂O) has been very successful in overcoming some of the problems associated with nitridation in ammonia. The authors have observed that the extent of N₂O oxidation has a strong detrimental effect on the drain leakage current of MOS transistors in the off state. This phenomenon has been identified to be caused by an increase in the active area junction leakage current     

Crystallization and melting behaviour of poly(butylene terephthalate) in poly(butylene terephthalate)/polyarylate blends

Summary The crystallization and melting behaviour of poly(butylene terephthalate) has been studied in the pure state and in its blends with a polyarylate of bisphenol A and isophthalic/terephthalic acids. Differential scanning calorimetry has been used as experimental technique and the effects of different thermal treatments have been analyzed. Results show the hindrance for the crystallization of poly(butylene terephthalate) imposed by the presence of polyarylate, as well as the existence of multiple melting after isothermal crystallization. Explanations are given for the observed behaviours.