Institute Computacao


Institute Computacao

Time filter
Source Type

Rios E.,UESPI | Coelho I.M.,State University of Rio de Janeiro | Ochi L.S.,Institute Computacao | Boeres C.,Institute Computacao | Farias R.,Federal University of Rio de Janeiro
Proceedings - 28th IEEE International Symposium on Computer Architecture and High Performance Computing Workshops, SBAC-PADW 2016 | Year: 2016

In combinatorial optimization problems, the neighborhood search (NS) is a fundamental component for local search based heuristics. It consists of selecting a solution from a high cardinality set of neighbor solutions, by means of operations called moves. To perform this search, NS algorithms usually adopt two main approaches: selecting the first or best improving move. The Multi Improvement (MI) strategy is a recently proposed method that consists in exploring simultaneously multiple move operations during the NS phase aiming to reach good quality solutions with shorter computational steps. This paper presents a benchmark for MI strategies in hybrid CPU/GPU systems. This technique efficiently explores the CPU processing power together with the massive parallelism achieved by modern GPUs, emerging as an efficient alternative for classic CPU neighborhood search strategies. The advantage of this approach depends heavily on finding the best tradeoff between CPU and GPU processing, as well as minimizing the memory transfers involved in the process. In the experiments, several MI configurations were tested in a hybrid CPU/GPU environment presenting better results than classical neighborhood search strategies for the Minimum Latency Problem, a hard combinatorial optimization problem. © 2016 IEEE.

Vieira-Damiani G.,University of Campinas | Ferro D.P.,University of Campinas | Adam R.L.,Institute Computacao | De Thomaz A.A.,Institute Fisica Gleb Wataghin | And 3 more authors.
Progress in Biomedical Optics and Imaging - Proceedings of SPIE | Year: 2011

Elastic and collagen fibers are essential components of the aorta, the remodeling of these structures is accompanied with aging in various diseases and life-threatening events. While the elastic fibers confer resilience to major blood vessels collagen confers resistance to the same. Elastic fibers are easily visualized in the fluorescent light when stained with hematoxylin eosin. Second Harmonic Generation (SHG) is a non linear signal that occurs only in molecules without inversion symmetry and is particularly strong in the collagen fibers arranged in triple helices. The aim of this paper is to describe the distribution of collagen in the thickness of the thoracic aorta, and to demonstrate the distribution of between elastic fibers. The images were acquired in a multifoton microscopy and both signals, Two-phtoton excitaded fluorescence (TPEF) and SHG, were excited by a Ti:Sapphire laser. We used a band pass filter to filter the SHG signal from the TPEF signal. The thickness of the aorta varies 2-3 mm, and the image was composed of the juxtaposition of images of 220 × 220 microns. We acquired images of a histological slide of the thoracic aorta stained with picrosirius red (specific for collagen) at a wavelength of 670nm SHG subsequently acquired images with the same region and observed that the images are overlapping. Therefore, the following images were acquired by confocal microscopy (fluorescence of eosin for visualization of elastic fibers) and for collagen SHG.After reconstruction of the images, we observed the distribution of collagen along the aorta. © 2011 SPIE.

Amaral M.A.,Federal Technological University of Paraná | Bim S.A.,Federal Technological University of Paraná | Boscarioli C.,West Parana State University | Maciel C.,Institute Computacao
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2015

The participation of women in workgroups is essential to any kind of job. Nowadays Computer Science has little presence of women. Consequently, it is necessary to motivate girls to choose Computing as a career. This paper describes an experiment consisting of Human-Computer Interaction (HCI) activities that were used to demonstrate that Computing is more than only programming. The activities were based on the concepts and materials of the Computer Science Unplugged Project. Besides, this work presents the results with the HCI activities performed in the context two Brazilian projects. The results are qualitative in nature and provide insights about HCI activities to elementary school student girls. Fifty-two girls from two different cities and four different public schools participated. All students developed the five tasks designed by CSU materials and were able to articulate concepts of HCI. © Springer International Publishing Switzerland 2015.

De Mendonca P.G.A.,Institute Educacao | Maciel C.,Institute Educacao | Maciel C.,Institute Computacao | Viterbo J.,Institute Computacao
Information Polity | Year: 2015

Publishing open government data and providing visualizations that favor the quick understanding and interpretation of the associated information is a task that requires systematization and the use of various technologies. Due to technical limitations (dependence on technological platforms, for example), this is not a standard practice among government system developers. In this work, we present a case study that involves the creation of a map to visualize information about the infestation of Aedes aegypti (dengue vector) in the municipality of Cuiabá, based on the selection and analysis of open data. This case study not only demonstrates that offering an easy way for presenting information on public health brings benefits to the community, but it also discusses a strategy for government agencies publishing open data and creating visualizations that combine information from various sources. © 2015 - IOS Press and the authors.

Queiroz F.,Institute Computacao | Walter M.,Federal University of Rio Grande do Sul
Brazilian Symposium of Computer Graphic and Image Processing | Year: 2012

The seamless integration of the shape and visual attributes of virtual objects is still one of the greatest challenges in Computer Graphics. For some natural objects, such as patterned animals, shape and appearance are mutually connected and therefore the individual treatment of these two aspects difficult the whole process and limits the visual results. One approach to solve this problem is to create shape and appearance together, thus generating so-called intelligent textures, since they can adapt to the surface of the object according to geometric information. The Clonal Mosaic model presented an approach for intelligent texturing of fur patterns seen in some mammals, particularly the big cats and giraffe. This paper extends this model to account for biologically plausible contrasting fur patterns, mostly seen in black and white, either regular - as seen in zebras, or irregular - as seen in cows and horses, among other animals. The main contributions of this work are the addition of a neural crest model, local control for parameters, and also vector field definition on the object's surface for simulation control. The results synthesized for various mammals with contrasting patterns such as cows, horses, and zebras, and other contrasting patterns found in frogs, for example, confirm the advantages of an integrated approach such as the one provided by the extended Clonal Mosaic procedural model. © 2012 IEEE.

De Souza C.S.,Pontifical Catholic University of Rio de Janeiro | Garcia A.C.B.,Institute Computacao | Slaviero C.,Institute Computacao | Pinto H.,Institute Computacao | Repenning A.,University of Colorado at Boulder
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011

Computational thinking involves many different abilities, including being able to represent real and imaginary worlds in highly constrained computer languages. These typically support very selective kinds of perspectives, abstractions and articulation compared to the unlimited possibilities provided by natural languages. This paper reports findings from a qualitative empirical study with novice programmers, carried out with AgentSheets in a Brazilian public school. The driving research question was: How do meanings expressed in natural language narratives relate to computational constructs expressed in programs produced by novices? We used semiotic and linguistic analysis to compare meaning representations in natural and artificial texts (game descriptions in Brazilian Portuguese and Visual AgenTalk code). We looked for recurring relations and what they might mean in the context of computational thinking education. Our findings suggest that the semiotic richness of AgentSheets can be explored to introduce different aspects of computational thinking in principled and heoretically-informed ways. © 2011 Springer-Verlag.

Gomes R.D.V.A.,Pontifical Catholic University of Rio de Janeiro | Casanova M.A.,Pontifical Catholic University of Rio de Janeiro | Lopes G.R.,Pontifical Catholic University of Rio de Janeiro | Paes Leme L.A.P.,Institute Computacao
Lecture Notes in Business Information Processing | Year: 2015

The Linked Data best practices recommend to publish a new tripleset using well-known ontologies and to interlink the new tripleset with other triplesets. However, both are difficult tasks. This paper describes CRAWLER-LD, a metadata crawler that helps selecting ontologies and triplesets to be used, respectively, in the publication and the interlinking processes. The publisher of the new tripleset first selects a set T of terms that describe the application domain of interest. Then, he submits T to CRAWLER-LD, which searches for triplesets whose vocabularies include terms direct or transitively related to those in T. CRAWLER-LD returns a list of ontologies to be used for publishing the new tripleset, as well as a list of triplesets that the new tripleset can be interlinked with. CRAWLER-LD focuses on specific metadata properties, including subclass of, and returns only metadata, hence the classification “metadata focused crawler”. © Springer International Publishing Switzerland 2015.

de Barros Santos J.,Pontifical Catholic University of Rio de Janeiro | Lopes Vieira B.,Institute Computacao | Haeusler E.H.,Pontifical Catholic University of Rio de Janeiro
Electronic Notes in Theoretical Computer Science | Year: 2016

This paper presents results on the definition of a sequent calculus for Minimal Implicational Propositional Logic (M→) aimed to be used for provability and counter-model generation in this logic. The system tracks the attempts to construct a proof in such a way that, if the original formula is a M→ tautology, the tree structure produced by the proving process is a proof, otherwise, it is used to construct a counter-model using Kripke semantics. © 2016 The Author(s)

Mestria M.,Institute Computacao | Satoru Ochi L.,Institute Computacao | De Lima Martins S.,Institute Computacao
Computers and Operations Research | Year: 2013

In this paper, we propose new heuristics using several path-relinking strategies to solve the Clustered Traveling Salesman Problem (CTSP). The CTSP is a generalization of the Traveling Salesman Problem (TSP) in which the set of vertices is partitioned into clusters and the objective is to find a minimum cost Hamiltonian cycle such that the vertices of each cluster are visited continuously. A comparison among the performance of the several different adopted path-relinking strategies is presented using instances with up to 2000 vertices and clusters varying between 4 and 150 vertices. Also computational experiments were performed to compare the performance of the proposed heuristics with an exact algorithm and a Genetic Algorithm. The obtained computational results showed that the proposed heuristics were able to obtain competitive results related to the quality of the solutions and computational execution time. © 2012 Elsevier Ltd. All rights reserved.

Muriana L.M.,Institute Computacao | Maciel C.,Institute Of Computacao | Garcia A.C.B.,Institute Computacao
CEUR Workshop Proceedings | Year: 2013

In June 2013, Brazil witnessed several manifestations different from anything previously seen, since president Collor's impeachment. However, there is a big difference: now they are organized in few days. With Facebook, physical and virtual communities spread fast and easily organize demonstations. Popular mobilization for mutual concern issues, supported by technology, allowed the gathering of information by collective intelligence. The fanpage of the demonstration and that of the demonstration events in Facebook are studied in this paper, in order to analyze the interaction among users and to identify if this social networking can be considered a virtual community. Copyright © by the paper's authors.

Loading Institute Computacao collaborators
Loading Institute Computacao collaborators