F-MPJ: scalable Java message-passing communications on parallel systems

This paper presents F-MPJ (Fast MPJ), a scalable and efficient Message-Passing in Java (MPJ) communication middleware for parallel computing. The increasing interest in Java as the programming language of the multi-core era demands scalable performance on hybrid architectures (with both shared and distributed memory spaces). However, current Java communication middleware lacks efficient communication support. F-MPJ boosts this situation by: (1) providing efficient non-blocking communication, which allows communication overlapping and thus scalable performance; (2) taking advantage of shared memory systems and high-performance networks through the use of our high-performance Java sockets implementation (named JFS, Java Fast Sockets); (3) avoiding the use of communication buffers; and (4) optimizing MPJ collective primitives. Thus, F-MPJ significantly improves the scalability of current MPJ implementations. A performance evaluation on an InfiniBand multi-core cluster has shown that F-MPJ communication primitives outperform representative MPJ libraries up to 60 times. Furthermore, the use of F-MPJ in communication-intensive MPJ codes has increased their performance up to seven times.     

A Geometric Algebra Model for the Image Formation Process of Paracatadioptric Cameras

Omnidirectional vision sensors capture a wide field of view than can benefit many robotic applications. One type of omnidirectional vision sensor is the paracatadioptric. A paracatadioptric sensor combines a parabolic mirror and a camera inducing an orthographic projection. This combination provides a wide field of view while maintaining the single center of projection which is a desirable property of these sensors. Furthermore, lines are projected as circles on the paracatadioptric image plane. In contrast with traditional perspective cameras the image formation process of paracatadioptric sensors is no longer linear. However in this paper we present a model which is able to linearize it. This linearization is based on the fact that the paracatadioptric projection can be represented by a sphere inversion, that belongs to the conformal group Rⁿ\mathcal{R}^{n} which is isomorphic to the Lorentz group in Rⁿ⁺¹\mathcal{R}^{n+1}. Thus a nonlinear conformal transformation can be represented with an equivalent linear Lorenz transformation, which can be represented as a versor in the CGA. Therefore the present model can be applied algebraically not only to points, but also to point-pairs, lines, circles in the same way to all them and in a linear form. The benefits of the proposed method will be reflected on the development of complex applications that use paracatadioptric sensors.     

Towards a unified framework for opinion retrieval, mining and summarization

The exponential increase of subjective, user-generated content since the birth of the Social Web, has led to the necessity of developing automatic text processing systems able to extract, process and present relevant knowledge. In this paper, we tackle the Opinion Retrieval, Mining and Summarization task, by proposing a unified framework, composed of three crucial components (information retrieval, opinion mining and text summarization) that allow the retrieval, classification and summarization of subjective information. An extensive analysis is conducted, where different configurations of the framework are suggested and analyzed, in order to determine which is the best one, and under which conditions. The evaluation carried out and the results obtained show the appropriateness of the individual components, as well as the framework as a whole. By achieving an improvement over 10% compared to the state-of-the-art approaches in the context of blogs, we can conclude that subjective text can be efficiently dealt with by means of our proposed framework.     

Eigenclassifiers for combining correlated classifiers

In practice, classifiers in an ensemble are not independent. This paper is the continuation of our previous work on ensemble subset selection [A. Ula?, M. Semerci, O.T. Y?ld?z, E. Alpayd?n, Incremental construction of classifier and discriminant ensembles, Information Sciences, 179 (9) (2009) 1298–1318] and has two parts: first, we investigate the effect of four factors on correlation: (i) algorithms used for training, (ii) hyperparameters of the algorithms, (iii) resampled training sets, (iv) input feature subsets. Simulations using 14 classifiers on 38 data sets indicate that hyperparameters and overlapping training sets have higher effect on positive correlation than features and algorithms. Second, we propose postprocessing before fusing using principal component analysis (PCA) to form uncorrelated eigenclassifiers from a set of correlated experts. Combining the information from all classifiers may be better than subset selection where some base classifiers are pruned before combination, because using all allows redundancy.     

Cost-conscious comparison of supervised learning algorithms over multiple data sets

In the literature, there exist statistical tests to compare supervised learning algorithms on multiple data sets in terms of accuracy but they do not always generate an ordering. We propose Multi²Test, a generalization of our previous work, for ordering multiple learning algorithms on multiple data sets from “best” to “worst” where our goodness measure is composed of a prior cost term additional to generalization error. Our simulations show that Multi²Test generates orderings using pairwise tests on error and different types of cost using time and space complexity of the learning algorithms.     

Distance maps from unthresholded magnitudes

A straightforward algorithm that computes distance maps from unthresholded magnitudes is presented, suitable for still images and video sequences. While results on binary images are similar to classic Euclidean Distance Transforms, the proposed approach does not require a binarization step. Thus, no thresholds are needed and no information is lost in intermediate classification stages. Experiments include the evaluation of spatial and temporal coherence of distance map values, showing better results in both measurements than those obtained with Sobel or Deriche gradients and classic chessboard distance transforms.     

A case study on business process recovery using an e‐government system

Business processes have become one of the key assets of organization, since these processes allow them to discover and control what occurs in their environments, with information systems automating most of an organization's processes. Unfortunately, and as a result of uncontrolled maintenance, information systems age over time until it is necessary to replace them with new and modernized systems. However, while systems are aging, meaningful business knowledge that is not present in any of the organization's other assets gradually becomes embedded in them. The preservation of this knowledge through the recovery of the underlying business processes is, therefore, a critical problem. This paper provides, as a solution to the aforementioned problem, a model‐driven procedure for recovering business processes from legacy information systems. The procedure proposes a set of models at different abstraction levels, along with the model transformations between them. The paper also provides a supporting tool, which facilitates its adoption. Moreover, a real‐life case study concerning an e‐government system applies the proposed recovery procedure to validate its effectiveness and efficiency. The case study was carried out by following a formal protocol to improve its rigor and replicability. Copyright © 2011 John Wiley & Sons, Ltd.     

A model-driven process for the modernization of component-based systems

Software modernization is critical for organizations that need cost-effective solutions to deal with the rapid obsolescence of software and the increasing demand for new functionality. This paper presents the XIRUP modernization methodology, which proposes a highly iterative process, structured into four phases: preliminary evaluation, understanding, building and migration. This modernization process is feature-driven, component-based, focused on the early elicitation of key information, and relies on a model-driven approach with extensive use of experience from the previous projects. XIRUP has been defined in the European IST project MOMOCS, which has also built a suite of support tools. This paper introduces the process using a case study that illustrates its activities, related tools and results. The discussion highlights the specific characteristics of modernization projects and how a customized methodology can take advantage of them.