The hierarchical topology optimization of three-dimensional structures is addressed in this paper. The structure lay-out (macroscale) and material microstructure (microscale) are optimized concurrently to minimize compliance. It is assumed that structure is made up of a periodic cellular material. This two-scale optimization problem is subjected to global/local material resource constraints and to specific local constraints controlling the perimeter on material microstructure design. To handle local constraints a proper algorithmic strategy is used combining two versions of MMA (Method of Moving Asymptotes). Local design problems can be solved independently and one takes advantage of this property by using parallel computing techniques to speed-up the solution task. Concerning applications one might think on man's made structures but the model is used to gain insight in natural hierarchical materials such as bone. At apparent level (macroscale) bone exhibits two types of tissues in a well-defined arrangement (cortical and trabecular) while going down on length scale (microscale) one distinguish trabeculae from marrow space. Actually, bone adapts to the mechanical environment by changing its microstructure and this natural adaptation is simulated in the present work. Further, among morphometric parameters used to characterize trabecular architecture there is the surface-to-volume ratio, which depends on anatomical site of the trabecular region. This biological feature of bone architecture is taken into account in the present study by means of a local material design constraint involving perimeter control. The results obtained for a proximal femur bone show a fine agreement with real bone. 2.

This paper compares the accuracy of combined classifiers in medical data bases to the same knowledge discovery techniques applied to generic data bases. Specifically, we apply Bagging and Boosting methods for 16 medical and 16 generic data bases and compare the accuracy results with a more traditional approach (C4.5 algorithm). Bagging and Boosting methods are applied using different numbers of classifiers and the accuracy is computed using a cross-validation technique. This paper main contribution resides in recommend the most accurate method and possible parameterization for medical data bases and an initial identification of some characteristics that make medical data bases different from generic ones.

The use of ensemble classifiers, e.g., Bagging and Boosting, is wide spread to machine learning. However, most of studies in this area are based on empirical comparisons that suffer from a lack of care to the randomness of these methods. This paper describes the dangers of experiments with ensemble classifiers by analyzing the efficiency of Bagging and Boosting methods over 32 different data sets. The experiments show that variations due to randomness are often more relevant than the advantages among methods encountered in the literature. This paper main contribution is the claim, supported by statistical analysis, that no empirical comparison of ensemble classifiers can be scientifically done without paying attention to the random choices taken.

4710-057 Braga -Portugal Portable Ultrasound Pachymetry: Accuracy and Repeatability 2 ABSTRACT Purpose: To assess the accuracy and repeatability of central corneal thickness measurements taken with a new portable ultrasound pachymeter. Methods: Central thickness measurements were taken with a portable and a conventional ultrasound pachymeters in 57 right corneas of fifty-seven young adults (19 males, 38 females) aged 18 to 44 years (mean ± SD, 22.95 ± 3.92). Three repeated measures were obtained and then compared to obtain the repeatability of each instrument and the agreement between both pachymeters. The three readings taken with the portable pachymeter were compared against each other in order to evaluate intra-session repeatability and bias of each individual measurement regarding the mean of three.

A construção de ontologias é um processo complexo que compreende etapas como a extração de conceitos de domínio e a extração de relações taxonômicas e não-taxonômicas entre esses conceitos. A etapa de extração de relações não-taxonômicas é a mais negligenciada, especialmente para textos em língua portuguesa. Com isto, este trabalho apresenta uma proposta de extração de relações não-taxonômicas a partir de textos em português representados por uma lista de conceitos e informações contextuais automaticamente extraídos pela ferramenta ExATOlp.

This paper proposes a technique to build entity profiles starting from a set of defining corpora, i.e., a corpus considered as the definition of each entity. The proposed technique is applied in a classification task in order to determine how much a text, or corpus, is related to each of the profiled entities. This technique is general enough to be applied to any kind of entity, however, this paper experiments are conduct over entities describing a set of professors of a computer science graduate school through their advised M.Sc. thesis and Ph.D. dissertations. The profiles of each entity are applied to categorize other texts into one of the builded profiles. The analysis of the obtained results illustrates the power of the proposed technique.

Measuring productivity in globally distributed projects is crucial to improve team performance. These measures often display information on whether a given project is moving forward or starts to demonstrate undesired behaviors. In this paper we are interested in showing how analytical models could deliver insights for the behavior of specific distributed software collaboration projects. We present a model for distributed projects using stochastic automata networks (SAN) formalism to estimate, for instance, the required level of coordination for specific project configurations. We focus our attention on the level of interaction among project participants and its close relation with team’s productivity. The models are parameterized for different scenarios and solved using numerical methods to obtain exact solutions. We vary the team’s expertise and support levels to measure the impact on the overall project performance. As results, we present our derived productivity index for all sce...

O presente artigo intenta apresentar um projeto de interface entre Linguística de Corpus e Processamento de Linguagem Natural (PLN) em andamento na Pontifícia Universidade Católica do RS. Tal projeto consiste na exploração, via corpus escrito, de teses e dissertações em Linguística oriundas do Programa de Pós-Graduação em Letras da referida instituição de ensino superior, por meio de um software chamado ExATO (LOPES, 2012). A partir dessa ferramenta produzida em PLN, podemos ter em mãos uma série de recursos linguísticos que permitem a prossecução da nossa análise, tais como, a detecção de Hierarquia de Conceitos, Listas de Termos Extraídos, Concordanciador de Termos e Nuvens de Conceitos a fim de embasarmos uma proposta de ensino de Português como Língua Adicional, com vistas à proficiência dentro do gênero acadêmico por alunos não-falantes de Língua Portuguesa. Trazemos uma série de resultados já disponibilizados pelo ExATO, os quais fomentam a discussão acerca dos temas visualiza...