WHAD: Wikipedia historical attributes data

This paper describes the generation of temporally anchored infobox attribute data from the Wikipedia history of revisions. By mining (attribute, value) pairs from the revision history of the English Wikipedia we are able to collect a comprehensive knowledge base that contains data on how attributes change over time. When dealing with the Wikipedia edit history, vandalic and erroneous edits are a concern for data quality. We present a study of vandalism identification in Wikipedia edits that uses only features from the infoboxes, and show that we can obtain, on this dataset, an accuracy comparable to a state-of-the-art vandalism identification method that is based on the whole article. Finally, we discuss different characteristics of the extracted dataset, which we make available for further study.     

Semi-parametric nonlinear regression and transformation using functional networks

Functional networks are used to solve some nonlinear regression problems. One particular problem is how to find the optimal transformations of the response and/or the explanatory variables and obtain the best possible functional relation between the response and predictor variables. After a brief introduction to functional networks, two specific transformation models based on functional networks are proposed. Unlike in neural networks, where the selection of the network topology is arbitrary, the selection of the initial topology of a functional network is problem driven. This important feature of functional networks is illustrated for each of the two proposed models. An equivalent, but simpler network may be obtained from the initial topology using functional equations. The resultant model is then checked for uniqueness of representation. When the functions specified by the transformations are unknown in form, families of linear independent functions are used as approximations. Two different parametric criteria are used for learning these functions: the constrained least squares and the maximum canonical correlation. Model selection criteria are used to avoid the problem of overfitting. Finally, performance of the proposed method are assessed and compared to other methods using a simulation study as well as several real-life data.     

Recursive identification of switched ARX systems

We consider the problem of recursively identifying the parameters of a deterministic discrete-time Switched Auto-Regressive eXogenous (SARX) model, under the assumption that the number of models, the model orders and the mode sequence are unknown. The key to our approach is to view the identification of multiple ARX models as the identification of a single, though more complex, lifted dynamical model built by applying a polynomial embedding to the input/output data. We show that the dynamics of this lifted model do not depend on the value of the discrete state or the switching mechanism, and are linear on the so-called hybrid model parameters. Therefore, one can identify the parameters of the lifted model using a standard recursive identifier applied to the embedded input/output data. The estimated hybrid model parameters are then used to build a polynomial whose derivatives at a regressor give an estimate of the parameters of the ARX model generating that regressor. The estimated ARX model parameters are shown to converge exponentially to their true values under a suitable persistence of excitation condition on a projection of the embedded input/output data. Such a condition is a natural generalization of the well known result for ARX models. Although our algorithm is designed for perfect input/output data, our experiments also evaluate its performance as a function of the level of noise for different choices of the number of models and model orders. We also present an application to temporal video segmentation.     

On improving the performance of large parallel lattice Boltzmann flow simulations in heterogeneous porous media

Classical Cartesian domain decompositions for parallel lattice Boltzmann simulations of fluid flow through heterogeneous porous media are doomed to workload imbalance as the number of processors increases, thus leading to decreasing parallel performance. A one-lattice lattice Boltzmann method (LBM) implementation with vector data structure combined with even fluid node partitioning domain decomposition and fully-optimized data transfer layout is presented. It is found to provide nearly-optimal workload balance, lower memory usage and better computational performance than classical slice decomposition techniques using sparse matrix data structures. Predictive memory usage and parallel performance models are also established and observed to be in very good agreement with data corresponding to numerical fluid flow simulations performed through 3-dimensional packings of cylinders and polydisperse spheres.     

Evolutionary Learning of Dynamic Naive Bayesian Classifiers

Many problems such as voice recognition, speech recognition and many other tasks have been tackled with Hidden Markov Models (HMMs). These problems can also be dealt with an extension of the Naive Bayesian Classifier (NBC) known as Dynamic NBC (DNBC). From a dynamic Bayesian network (DBN) perspective, in a DNBC at each time there is a NBC. NBCs work well in data sets with independent attributes. However, they perform poorly when the attributes are dependent or when there are one or more irrelevant attributes which are dependent of some relevant ones. Therefore, to increase this classifier accuracy, we need a method to design network structures that can capture the dependencies and get rid of irrelevant attributes. Furthermore, when we deal with dynamical processes there are temporal relations that should be considered in the network design. In order to learn automatically these models from data and increase the classifier accuracy we propose an evolutionary optimization algorithm to solve this design problem. We introduce a new encoding scheme and new genetic operators which are natural extensions of previously proposed encoding and operators for grouping problems. The design methodology is applied to solve the recognition problem for nine hand gestures. Experimental results show that the evolved network has higher average classification accuracy than the basic DNBC and a HMM.     

Real-time massive convolution for audio applications on GPU

Massive convolution is the basic operation in multichannel acoustic signal processing. This field has experienced a major development in recent years. One reason for this has been the increase in the number of sound sources used in playback applications available to users. Another reason is the growing need to incorporate new effects and to improve the hearing experience. Massive convolution requires high computing capacity. GPUs offer the possibility of parallelizing these operations. This allows us to obtain the processing result in much shorter time and to free up CPU resources. One important aspect lies in the possibility of overlapping the transfer of data from CPU to GPU and vice versa with the computation, in order to carry out real-time applications. Thus, a synthesis of 3D sound scenes could be achieved with only a peer-to-peer music streaming environment using a simple GPU in your computer, while the CPU in the computer is being used for other tasks. Nowadays, these effects are obtained in theaters or funfairs at a very high cost, requiring a large quantity of resources. Thus, our work focuses on two mains points: to describe an efficient massive convolution implementation and to incorporate this task to real-time multichannel-sound applications.     

A historical perspective of algorithmic lateral inhibition and accumulative computation in computer vision

Certainly, one of the prominent ideas of Professor José Mira was that it is absolutely mandatory to specify the mechanisms and/or processes underlying each task and inference mentioned in an architecture in order to make operational that architecture. The conjecture of the last fifteen years of joint research has been that any bottom-up organization may be made operational using two biologically inspired methods called “algorithmic lateral inhibition”, a generalization of lateral inhibition anatomical circuits, and “accumulative computation”, a working memory related to the temporal evolution of the membrane potential. This paper is dedicated to the computational formulation of both methods. Finally, all of the works of our group related to this methodological approximation are mentioned and summarized, showing that all of them support the validity of this approximation.     

A formal analysis of database replication protocols with SI replicas and crash failures

This paper provides a formal specification and proof of correctness of a basic Generalized Snapshot Isolation certification-based data replication protocol for database middleware architectures. It has been modeled using a state transition system, as well as the main system components, allowing a perfect match with the usual deployment in a middleware system. The proof encompasses both safety and liveness properties, as it is commonly done for a distributed algorithm. Furthermore, a crash failure model has been assumed for the correctness proof, although recovery analysis is not the aim of this paper. This allows an easy extension toward a crash-recovery model support in future works. The liveness proof focuses in the uniform commit: if a site has committed a transaction, the rest of sites will either commit it or it would have crashed.     

Condition-based diagnosis of mechatronic systems using a fractional calculus approach

While fractional calculus (FC) is as old as integer calculus, its application has been mainly restricted to mathematics. However, many real systems are better described using FC equations than with integer models. FC is a suitable tool for describing systems characterised by their fractal nature, long-term memory and chaotic behaviour. It is a promising methodology for failure analysis and modelling, since the behaviour of a failing system depends on factors that increase the model’s complexity. This paper explores the proficiency of FC in modelling complex behaviour by tuning only a few parameters. This work proposes a novel two-step strategy for diagnosis, first modelling common failure conditions and, second, by comparing these models with real machine signals and using the difference to feed a computational classifier. Our proposal is validated using an electrical motor coupled with a mechanical gear reducer.