CLEI2015 XLI Conferencia Latinoamericana de Informa tica ... · i Abstracts of the XLI Conferencia Latinoamericana de Informatica Arequipa, Peru , Octubre/Outubro/October 19th-23rd

CLEI2015

XLI Conferencia Latinoamericana de Informaticahttp://clei.org/clei2015

Octubre/Outubro/October 19th-23rd, 2015, Arequipa, PeruHosted by: SPC, UNSA, UCSM, UCSP, ULS

Editors:Hector Cancela

Alex Cuadros-VargasErnesto Cuadros-Vargas

2

Abstracts

XLI Conferencia Latinoamericana de Informaticahttp://clei.org/clei2015

CLEI2015

Arequipa, Octubre/Outubro/October 19th-23rd,2015

3

Local Organizers:

Sociedad Peruana de Computacion

Universidad Nacional de San Agustın, Arequipa

Universidad Catolica de Santa Marıa, Arequipa

Universidad Catolica San Pablo, Arequipa

Universidad La Salle, Arequipa

Universidad Nacional de Ingenierıa, Lima

Universidad Nacional Mayor de San Marcos, Lima

Universidad Ricardo Palma, Lima

Universidad Nacional San Cristobal de Huamanga, Ayacucho

Universidad Nacional del Altiplano, Puno

Universidad Nacional Micaela Bastidas, Abancay

Asociacion Peruana de Productores de Software (APESOFT)

EDDAS, Arequipa

IEEE Seccion Peru

ACM UCSP Chapter

Silver Supporters:

UPGRADE, Peru

Oracle Academy

i

Abstracts of the XLI Conferencia Latinoamericana de InformaticaArequipa, Peru, Octubre/Outubro/October 19th-23rd

General ChairsHector Cancela, [email protected] Cuadros Vargas, [email protected]

Latin American Symposium on Computer Graphics, Virtual Reality, and Image Pro-cessing,Juan Carlos Gutierrez, [email protected] Gustavo Nonato, [email protected]

Latin American Symposium on Software EngineeringAdenilso da Silva Simao, [email protected] Condori-Fernandez, [email protected]

Latin American Symposium on Informatics and SocietyAlvaro Fernandez, [email protected] Tansini, [email protected]

Latin American Symposium on Operational Research and Artificial IntelligenceDennis Barrios Aranibar, [email protected] Pinto, [email protected]

Latin American Symposium on Infrastructure, Hardware, and SoftwareYvan Jesus Tupac Valdivia, [email protected] Tirado, [email protected]

Latin American Symposium on Large Scale Information SystemsGuillermo Calderon-Ruiz, [email protected] Astudillo, [email protected]

Latin American Symposium on Data and Information ManagementJorge Perez Rojas, [email protected] Angles, [email protected]

Latin American Symposium on Theoretical Computer ScienceCristian Lopez Del Alamo, [email protected] Fernandez-Baca, Iowa State University, [email protected]

XXII Iberoamerican Congress on Higher Education in Computing (CIESC 2015)Tony Clear, [email protected] Moro, [email protected]

Chairs de los Eventos AsociadosXXI Latin American Contest of Master Thesis (CLTM 2015)Andre Ponce de Leon F. de Carvalho, [email protected] Motz, Uruguay, [email protected]

I Latin American Contest of PhD Thesis (CLTD 2015)Gabriela Marin, [email protected] Santos, [email protected]

The VI Latin American Women in Computing Congress (LAWCC 2015)Raquel Esperanza Patino Escarcina, [email protected] Delgado, [email protected]

The IV Workshop on Nomenclature and Accreditation of Computing ProgramsErnesto Cuadros-Vargas, [email protected] Sabiguero, [email protected]

ii

XLI Conferencia Latinoamericana de InformaticaOrganizing Committee

ChairErnesto Cuadros Vargas, San Pablo Catholic University, Peru. ([email protected])

Protocol and Culture CommitteeWilber Ramos, UNSA, Peru. ([email protected])

Travel Agency CommitteeEdgard Rodriguez, Agencia de Viajes Turismo Tropical SAC, Peru. (turismotropical [email protected])

Registration CommitteeGraciela Lecireth Meza Lovon, San Pablo Catholic University, Peru. ([email protected])

Logistic CommitteeMedardo Delgado, Universidad La Salle, Peru.

Infraestructure and Technology CommitteeJulio Guillermo Paredes, EDDAS, Peru. ([email protected])

Advertising CommitteeJose Herrera, Universidad Nacional de San Agustin, Peru. ([email protected])Romilda Talavera Eguiluz ([email protected])

Assistance for GroupsJose Luis Montoya Vargas, San Pablo Catholic University, Peru. ([email protected])

Support for attendeesJose Luis Montoya Vargas, San Pablo Catholic University, Peru. ([email protected])Olha Sharhorodska, Universidad Nacional de San Agustın, Peru. ([email protected])

Treasury CommitteeYvan Jesus Tupac Valdivia, San Pablo Catholic University, Peru. ([email protected])

iii

Message from Conference andGeneral Program ChairsCLEI (Centro Latinoamericano de Estudios en Informatica) is an organization composed of aproxi-matly ninety institutions granting Computing degrees from all countries in the Latin American region.

CLEI has promoted annual meetings on Computing in the Latin American region for more thanfour decades. These meetings promote the exchange of knowledge and progress in research on differentfields of Computing. The CLEI events have a high impact in the Latin American region, and alsoin other parts of the world. They have consolidated a number of symposia in different domains ofcomputing, and have created others in new research areas.

Throughout the years, the successful development of each meeting has enabled the CLEI organi-zation to consolidate a prestigious trajectory in the Computer Science community in Latin Americaas well as in other regions.

This is the second time Peru hosts the CLEI annual meeting. The first time was in 2004, both inArequipa. The present edition is a very special one, as it marks the 41st CLEI conference, and it is co-organized by more that 10 Peruvian institutions: Sociedad Peruana de Computacion UniversidadNacional de San Agustın (Arequipa) Universidad Catolica de Santa Marıa (Arequipa) UniversidadCatolica San Pablo (Arequipa) Universidad La Salle (Arequipa) Universidad Nacional de Ingenierıa(Lima) Universidad Nacional Mayor de San Marcos (Lima) Universidad Ricardo Palma (Lima)

Universidad Nacional San Cristobal de Huamanga (Ayacucho) Universidad Nacional del Altiplano(Puno) Universidad Nacional Micaela Bastidas (Abancay) Asociacion Peruana de Productores deSoftware (APESOFT) EDDAS (Arequipa) IEEE Peru Section and ACM UCSP Chapter showingthe cooperation and strenght of the Peruvian academic community.

For the last several years, the Latin American Computing Conference has been organized in a num-ber of symposia: 1) Latin American Symposium on Computer Graphics, Virtual Reality, and ImageProcessing, 2) Latin American Symposium on Software Engineering, 3) Latin American Symposiumon Informatics and Society, 4) Latin American Symposium on Operational Research and ArtificialIntelligence, and 5) Latin American Symposium on Infrastructure, Hardware, and Software, 6) LatinAmerican Symposium on Large Scale Information Systems, 7) Latin American Symposium on Da-ta and Information Management, 8) Latin American Symposium on Theoretical Computer Science.9) XXII Iberoamerican Symposium on Higher Education in Computing (SIESC 2015)

Additionally, CLEI 2015 includes the following associated events: 1) XXI Latin American MasterThesis Competition(CLTM 2015) 2) I Latin American PhD Thesis (CLTD 2015) 3) The VI LatinAmerican Women in Computing Congress (LAWCC 2015) 4) The IV Workshop on Nomenclature andAccreditation of Computing Programs

Over 400 contributions were received from different parts of America and Europe. All contributionswere subject to a rigorous review process, resulting in 120 accepted papers. The main proceedings ofthe conference are included and indexed in IEEE Xplore; at the same time, the best papers presentedat CLEI 2015 symposia and SIESC 2015 were invited to be published in the Electronic Notes inTheoretical Computer Science (ENTCS) journal and the CLEI Electronic Journal (CLEIej).

CLEI 2015 is sponsored by EDDAS, APESOFT and Oracle Academy. A conference like CLEI isonly possible by the involvement of a large community; we want to thank the CLEI steering committee,members of the program committees and the organizing committee, invited lectures and speakers,authors, teachers, students, administrative staff, and many other people who contributed towards thesuccess of the event.

Hector CancelaUniversidad de la Republica (Uruguay)

Chair of the Program Committee

Alex Cuadros-VargasUniversidad Catolica San Pablo (Peru)

Chair of the Program Committee

Ernesto Cuadros-VargasUniversidad Catolica San Pablo (Peru)

Chair of the Organizing Committee

v

IndexFull Program Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

Keynote . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Tutorial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Simposio Latinoamericano de Computacion Grafica, Realidad Virtual y Procesamiento de Imagenes(SLCGRVPI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Simposio Latinoamericano de Infraestructura, Hardware y Software (SLIHS) . . . . . . . . . . . . . . . . . . . . . 11

Simposio Latinoamericano de Investigacion de Operaciones e Inteligencia Artificial (SLIOIA) . . . . . 16

Simposio Latinoamericano de Informatica y Sociedad (SLIS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Simposio Latinoamericano de Sistemas de Informacion de Gran Escala (SLSIGE). . . . . . . . . . . . . . . . .27

Simposio Latinoamericano de Ingenierıa de Software (SLISW) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

Simposio Latinoamericano de Manejo de Datos e Informacion (SLMDI) . . . . . . . . . . . . . . . . . . . . . . . . . . 36

Simposio Latinoamericano de Teorıa Computacional (SLTC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

XXIII Simposio Iberoamericano de Educacion Superior en Computacion (SIESC). . . . . . . . . . . . . . . . .40

V Workshop en Nomenclatura y Acreditacion en Programas de Computacion (WNAPC) . . . . . . . . . 42

VII Congreso de la Mujer Latinoamericana en la Computacion (LAWCC) . . . . . . . . . . . . . . . . . . . . . . . . 43

XXII Concurso Latinoamericano de Tesis de Maestrıa (CLTM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

I Concurso Latinoamericano de Tesis de Doctorado (CLTD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

Collaboration Networks in Computing (Graduate). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48

Premios de Investigacion de Google para America Latina (Google) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

vi

General IndexOctober 19th (Lunes/Segunda feira/Monday)

Keynote , Sala/Room: A09:00-10:00 Next Generation of GPUs. Esteban Clua . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Simposio Latinoamericano de Informatica y SociedadSession: S1, Sala/Room: CChair: Francisco Mata (Costa Rica)

10:15-10:45 DICREVOA: A Proposal for the Design, Creation and Evaluation ofLearning Objects. Jorge J. Maldonado, Jorge L. Bermeo, Magali Mejıa . . . . . . . 21

10:45-11:15 Vulcanus: A Recommender System for Accessibility based on Trails. Is-mael Cardoso, Bruno Mota, Jorge Barbosa, Rodrigo Righi . . . . . . . . . . . . . . . . . . . . . . 21

11:15-11:45 Explorando o Conceito de Internet das Coisas Sociais em um AmbienteUniversitario utilizando NFC. Tiago Alves, Cristiano Andre da Costa, Rodrigoda Rosa Righi, Jorge Barbosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Simposio Latinoamericano de Investigacion de Operaciones e Inteligencia ArtificialSession: S1, Sala/Room: AChair: Diego Pinto Roa (Paraguay)

10:15-10:45 Optimizacion de Enjambre de Partıculas para Problemas de Muchos Ob-jetivos. Mateo Torres Bobadilla, Benjamın Baran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

10:45-11:15 HGVPRLB: a hybrid algorithm for solving binary problems. Josiane daCosta Vieira Rezende, Rone Ilıdio da Silva, Marcone Freitas Souza . . . . . . . . . . . . . 16

11:15-11:45 Active learning algorithms for multi-label data. Everton Cherman, GrigoriosTsoumakas, Maria Monard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Simposio Latinoamericano de Ingenierıa de SoftwareSession: Requirements engineering and risk management, Sala/Room: BChair: Nelly Condori-Fernandez (The Netherlands)

10:15-10:45 Mind Maps in the Requirements Traceability. Alessandro Silveira Duarte,Jose Augusto Fabri, Alexandre L’Erario, Elias Canhadas Genvigir . . . . . . . . . . . . . . 30

10:45-11:15 Use case technique for requirements modeling in distributed develop-ment environments. Viviana Alferillo, Maria Ines Lund, Gerardo Matturro . . . 30

11:15-11:45 A methodology to guide writing a Software Requirements Specificationdocument. Helcio A. Soares, Raimundo Moura . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

11:45-12:15 Collaborative Approach to Security Risk Management Information. Mai-con Balke, Lisandra Fontoura, Luis Alvaro de Lima Silva . . . . . . . . . . . . . . . . . . . . . . . 30

Simposio Latinoamericano de Infraestructura, Hardware y SoftwareSession: S1, Sala/Room: DChair: Jose Aguilar (Venezuela)

10:15-10:45 Scratchpad Memory Management Using Data-Prefetching. Ivan SaraivaSilva, Hildebrando Segundo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

10:45-11:15 Cloud Elasticity for HPC Applications: Observing Energy, Performanceand Cost. Vinicius Facco Rodrigues, Gustavo Rostirolla, Rodrigo da Rosa Righi,Cristiano Andre da Costa, Jorge Luis Victoria Barbosa . . . . . . . . . . . . . . . . . . . . . . . . . 11

11:15-11:45 Optimal Cloud Resource Allocation by Means of the Analytic HierarchyProcess. Valter Messias, Julio Estrella, Ricardo Ehlers . . . . . . . . . . . . . . . . . . . . . . . . . 11

11:45-12:15 A Multi-Objective Approach for Virtual Network Embedding. EnriqueDavalos Gimenez, Benjamın Baran, Cristian Aceval, Victor Franco . . . . . . . . . . . . . 11

Keynote , Sala/Room: A14:00-15:00 Los procesadores en el final de la ley de Moore. Francisco Tirado . . . . . . . . . 1

Simposio Latinoamericano de Ingenierıa de SoftwareSession: Software modeling, Sala/Room: BChair: Andrea Delgado (Uruguay)

vii

15:15-15:45 Evolution of a Model-driven Process Framework. Wilson Padua . . . . . . . . . 3115:45-16:15 An MDE-Based Graphical Tool for the Validation of MySQL Replica-

tion Models. Efraın Bautista Ubillus, Nora La Serna Palomino . . . . . . . . . . . . . . . . 3116:15-16:45 Dribbling Complexity in Model Driven Development Using Naked Ob-

jects, Domain Driven Design, and Software Design Patterns. Samuel A.Soares, Marcius Brandao, Mariela Cortes, Emmanuel Savio Silva Freire . . . . . . . . 31

16:45-17:15 Comparison of Software Process Models. A Systematic Literature Re-view. Christian Cano, Abraham Davila, Andres Melgar, Marcelo Pessoa . . . . . . . . 31

Simposio Latinoamericano de Investigacion de Operaciones e Inteligencia ArtificialSession: S2, Sala/Room: AChair: Dennis Barrios (Peru)

15:45-16:15 Conditional Monte Carlo with Intermediate Estimations for simulationof Markovian systems. Hector Cancela, Leslie Murray, Gerardo Rubino . . . . . . 16

16:15-16:45 A Tabu Search based heuristic for police units positioning. Nilson Mendes,Andre dos Santos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

16:45-17:15 Improving Subjectivity Detection for Spanish Texts using SubjectivityWord Sense Disambiguation based on Knowledge. Marco Sobrevilla Cabezu-do, Nora La Serna Palomino, Rolando Alberto Maguina Perez . . . . . . . . . . . . . . . . . . 17

Simposio Latinoamericano de Informatica y SociedadSession: S2, Sala/Room: CChair: Jairo Riano (Colombia)

15:45-16:15 Web Accessibility: Study web accessibility in public places of the Colom-bian State. Javier Antonio Ballesteros Ricaurte, Jairo Armando Riano Herrera21

16:15-16:45 Uso del Comercio Electronico para la Venta de Cafe Tostado de CostaRica: Estudio de Casos de Torrefactores de Cafe Costarricenses. FranciscoJ. Mata, Ariella Quesada . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

16:45-17:15 Emancipation of Access to Wheelwright’s Financial Information. Alejan-dro Sartorio, Guillermo Rodrıguez, Daniel Tedini, Marcelo Vaquero, Alejandro Her-nandez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

Simposio Latinoamericano de Infraestructura, Hardware y SoftwareSession: S2, Sala/Room: DChair: Diego Pinto (Paraguay)

15:45-16:15 An Experimental Study on the Effectiveness of Trilateration and Pro-babilistic Multilateration for Position Estimation in MANETs. AnabelPineda-Briseno, Rolando Menchaca-Mendez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

16:15-16:45 gr-isdbt: An ISDB-T 1-segment Receiver Implementation on GNU Ra-dio. Federico Larroca, Pablo Flores, Gabriel Gomez, Vıctor Gonzalez-Barbone, Pa-blo Belzarena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

16:45-17:15 MultiObjective Robust Network Design under Uncertain Traffic. An ap-proach based on Evolutionary Algorithms. Adolfo Arteta, Diego P. Pinto-Roa12

17:15-17:45 Auction-based Resource Provisioning in Cloud Computing. A Taxo-nomy. Sara Arevalos, Fabio Lopez-Pires, Benjamın Baran . . . . . . . . . . . . . . . . . . . . . . 13

Simposio Latinoamericano de Ingenierıa de SoftwareSession: Software Testing, Sala/Room: BChair: Adenilso Simao (Brazil)

viii

17:45-18:15 Using TDD for Developing Object-Oriented Software - A Case Study.Ramon Goncalves, Igor Lima, Heitor Costa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

18:15-18:45 Extracting Static and Dynamic Structural Information from Java Con-current Programs for Coverage Testing. Rafael R. Prado, Paulo S. L. Lopesde Souza, George G. M. Dourado, Simone R. S. Souza, Julio C. Estrella, Sarita M.Bruschi, Joao Lourenco . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

18:45-19:15 Automated Testing Framework for Mobile Applications based on User-Interaction Features and Historical Bug Information. Abel Mendez-Porras,Jorge Alfaro-Velasco, Marcelo Jenkins, Alexandra Martınez Porras . . . . . . . . . . . . . . 32


17:45-18:15 Logical Analysis of Multi-Class Data. Juan Avila-Herrera, Munevver MineSubasi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

18:15-18:45 Performance metrics in multi-objective optimization. Nery Riquelme,Christian Von Lucken, Benjamın Baran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

18:45-19:15 Semantic Mining in Clusters from Signaling Pathways Networks. CarlosRangel, Jose Aguilar, Junior Amilcar Altamiranda Perez . . . . . . . . . . . . . . . . . . . . . . . . 18

19:15-19:45 A Sanction-Application Mechanism considering Commitment Levels inHierarchical Organizations. Robert Junior, Emmanuel Freire, Mariela Cortes,Gustavo Campos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

Simposio Latinoamericano de Informatica y SociedadSession: S3, Sala/Room: CChair: Rodrigo Righi (Brazil)

17:45-18:15 PICTOAPRENDE: Application that contributes to the personal auto-nomy of children and youth with Autism Spectrum Disorder in Ecuador.Johanna Tobar, Danni De la Cruz, Paul Mejıa, Nancy Paredes, Andrea Cardenas,Edison Segovia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

18:15-18:45 Meneduca - Social School Network to Support the Educational Environ-ment. Guilherme Quirino, Nadav Mals, Vitor Groterhorst, Solange N. Alves deSouza, Luiz Sergio de Souza . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

18:45-19:15 Time series analysis of agro-meteorological data through algorithms in ascalable data mining case: Chili river watershed, Arequipa. Melissa Abar-ca, Karla Mariel Fernandez Fabian, Jose Herrera Quispe . . . . . . . . . . . . . . . . . . . . . . . .23

19:15-19:45 Cooperative estimation of Vehicular Traffic using Mobile Applications.Joaquın Lima, Ruben Lopez, Alfredo Campuzano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

Simposio Latinoamericano de Infraestructura, Hardware y SoftwareSession: S3, Sala/Room: DChair: Benjamın Baran (Paraguay)

17:45-18:15 GAIA Infrastructure: a Framework for the Management of Informationand Communication Technology Infrastructure. Luis Horacio Ramos Isique,Rodolfo Miranda de Barros, Bruno Bogaz Zarpelao . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13

18:15-18:45 Protection with Quality of Service in optical WDM networks usingMany-Objective Ant Colony Optimization. Benjamın Baran, Julio Paciello,Tania Nunez, Victor Ayala . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

18:45-19:15 The Use of Different Strategies of Search Space Reduction in Mitigationof Optimization Selection Problem. Nilton Queiroz Junior, Anderson Faustinoda Silva . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

19:15-19:45 Improved Batch Elimination: A Fast Algorithm to Identify and RemoveHarmful Compiler Optimizations. Ewerton Daniel de Lima, Anderson Faus-tino da Silva . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14

October 20th (Martes/Terca feira/Tuesday)

Keynote , Sala/Room: A

ix

09:00-10:00 Service Quality Assurance in Context-Aware Systems. Nelly Condori-Fernandez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Simposio Latinoamericano de Ingenierıa de SoftwareSession: Empirical software engineering, Sala/Room: BChair: Luca Cernuzzi (Paraguay)

10:15-10:45 An exploratory study about cross-project defect prediction: impact ofusing different classification algorithms and a measure of performance inbuilding predictive models.. Ricardo F. P. Satin, Igor Scaliante Wiese, Reginal-do Re . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

10:45-11:15 Avaliacao da Transparencia do Sistema de Compras do Governo Bra-siliero. Francisco Chaves Pinto, Denis Silva da Silveira, Jairo Simiao Dornelas,Charlie Silva Lopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

11:15-11:45 Graphical and Statistical Analysis of Software Evolution Using Couplingand Cohesion Metrics - An Exploratory Study. Raul Silva, Heitor Costa . 33

11:45-12:15 Performance and Accuracy conflict in Monitoring Tools for Web Servi-ces: a case study. Jael Zela Ruiz, Cecılia Rubira . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Simposio Latinoamericano de Infraestructura, Hardware y SoftwareSession: S4, Sala/Room: DChair: Yvan Tupac (Peru)

10:15-10:45 HTTP-WS-AD: An Anomaly Detector oriented to web applications andweb services. Jose Gimenez, Cristian Cappo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

10:45-11:15 Solving dense linear systems with hybrid ARM+GPU platforms. JuanPablo Silva, Ernesto Dufrechou, Enrique Quintana, Alfredo Remon, Peter Benner14

11:15-11:45 Network Virtualization in Optical Networks with Traffic Grooming. Enri-que Davalos Gimenez, Marcos Tilerıa, Aloysius Yu, Benjamın Baran . . . . . . . . . . . 15

11:45-12:15 A New Approach to the Massive Processing of Satellite Images. WilderNina, Alvaro Mamani-Aliaga, Rene Cruz, Juber Serrano, Jaime Cuba, Yoni Huay-nacho, Yessenia Yari, Pablo Yanyachi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Simposio Latinoamericano de Investigacion de Operaciones e Inteligencia ArtificialSession: S4, Sala/Room: AChair: Dennis Barrios (Peru)

10:15-10:45 Dynamic Approach for a Demand Responsive Transport Service. RenanJose dos Santos Viana, Andre Gustavo dos Santos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

10:45-11:15 Discrete Choquet integral based method for criteria synergy determi-nation. Ruben Bernal, Marcelo Karanik, Jose Ignacio Pelaez Sanchez, Estela delRocio Yanez Benavides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

11:15-11:45 A TrueSkill approach for movies recommendation. Laura Cruz Quispe, JoseEduardo Ochoa Luna . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

11:45-12:15 Bird Species Classification Using Spectrograms. Diego Rafael Lucio, YandreMaldonado e Gomes da Costa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Simposio Latinoamericano de Informatica y SociedadSession: S4, Sala/Room: CChair: Alvaro Fernandez (Peru)

10:15-10:45 Quality and maturity model for open data portals. Edgar Oviedo, Jose Nor-berto Mazon, Jose Jacobo Zubcoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

10:45-11:15 Cooperative Live Coding as an instructional model. Antonio de Carvalho Jr24

11:15-11:45 Predictive model of dengue focus applied to Geographic InformationSystems. Maximiliano Baez Gonzalez, Guillermo Gonzalez Rodas . . . . . . . . . . . . . . 24

11:45-12:15 Using the Kinect Sensor with Open Source Tools for the Developmentof Educational Games for Children on Pre-school Age. Romina Fernandez,Christian von Lucken . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Keynote , Sala/Room: A

x

14:00-15:00 Inteligencia Artificial en el Siglo XXI. Nicolas Kemper . . . . . . . . . . . . . . . . . . . . . .2

Simposio Latinoamericano de Informatica y SociedadSession: S5, Sala/Room: CChair: Solange Souza (Brazil)

15:15-15:45 Sumo Sensei: Design, Implementacao e Teste com Usuarios de uma Fe-rramenta Movel para Apoiar o Estudo de Kanjis Basicos. Fabio RochaMarques, Bruna de Menezes, Leonardo Cunha de Miranda, Erica Esteves Cunhade Miranda, Juvane Nunes Marciano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

15:45-16:15 Karuta Kanji: Jogo Educacional para Estudar e Praticar Vocabulariocom Kanjis da Lıngua Japonesa. Fabio Rocha Marques, Bruna de Menezes,Leonardo Cunha de Miranda, Erica Esteves Cunha de Miranda, Juvane NunesMarciano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

16:15-16:45 Katakana Star Samurai: A Mobile Tool to Support Learning of a basicJapanese Alphabet. Juvane Nunes Marciano, Jaime Oliveira, Bruna de Mene-zes, Leonardo Cunha de Miranda, Erica Esteves Cunha de Miranda . . . . . . . . . . . . . 25

16:45-17:15 Use of a OWL ontology for creating Interactive Learning Object. BrunoNogueira Luz, Rafael Santos, Andreza Silva Areao, Marcos Hideyuki Yokoyama,Luiz Egidio Cunha, Valeria Martins, Marcelo Guimaraes . . . . . . . . . . . . . . . . . . . . . . . 26

Simposio Latinoamericano de Teorıa ComputacionalSession: S1, Sala/Room: DChair: Cristian Lopez Del Alamo (Peru)

15:15-15:45 On discerning strings with finite automata. Abuzer Yakaryilmaz, J. AndresMontoya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

15:45-16:15 On the real-state processing of regular operations and The Sakoda-Sipser problem. J. Andres Montoya, David Casas . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38

16:15-16:45 A Model to Guide Dynamic Adaptation Planning in Self-Adaptive Sys-tems. Andres Paz, Hugo Arboleda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

16:45-17:15 Computing Translocation Distance by a Genetic Algorithm. Lucas da Sil-veira, Jose Luis Soncco-Alvarez, Thaynara A. de Lima, Mauricio Ayala-Rincon . 38

Simposio Latinoamericano de Ingenierıa de SoftwareSession: Software Measurement and Quality, Sala/Room: BChair: Adenilso Simao (Brazil)

15:15-15:45 How to Automatically Collect Oriented Object Metrics: A Study Basedon Systematic Review. Moshe Ribeiro, Rodrigo Reis, Antonio Abelem . . . . . . . . 34

15:45-16:15 Sistema de Analise de Incidentes para Melhoria Contınua. Lucas Filippi,Sandro Jose Rigo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

16:15-16:45 Factors driving the adoption of ISO / IEC 29110: a case study of a smallsoftware enterprise. Abraham Davila, Marcelo Schneck de Paula Pessoa . . . . . . 34

16:45-17:15 Mejora de la monitorizacion y ejecucion de procesos de negocio con inte-gracion y socializacion. Patricia Bazan, Jose Martinez Garro, Roxana Giandini,Javier Diaz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Tutorial , Sala/Room: A

15:15-19:15 Planificacion Optima de Sistemas Multinucleo conrestricciones de cali-dad de servicio, recursos y energıa. Rodrigo Santos . . . . . . . . . . . . . . . . . . . . . . . . . .6

VII Congreso de la Mujer Latinoamericana en la ComputacionSession: S1, Sala/Room: BChair: Andrea Delgado (Uruguay)

17:15-17:45 Participacao Feminina em Pesquisa na Plataforma Lattes no Brasil. Ma-ria Carolina Monard, Ana Maria Monteiro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

17:45-18:15 Digital Equity and Gender Issues in Latin America. Gabriela Marın . . . . . 4318:15-19:15 Acciones a futuro en Latinoamerica - discusion general. Andrea Delgado,

Raquel Patino . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4418:15-19:15 Acciones a futuro en Latinoamerica - discusion general. Andrea Delgado,

Raquel Patino . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

xi

V Workshop en Nomenclatura y Acreditacion en Programas de ComputacionSession: S1, Sala/Room: CChair: Ernesto Cuadros-Vargas (Peru)

17:15-17:45 Caso Colombia: La investigacion como factor determinante en el Ase-guramiento de la Calidad de las Instituciones de Educacion Superior.Richard Aroca Acosta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

17:45-18:15 Presentaciones de la problematica de diversos paıses de la region (par-te2). Ernesto Cuadros-Vargas, Ariel Sabiguero Yawelak . . . . . . . . . . . . . . . . . . . . . . . . .42



October 21st (Miercoles/Quarta feira/Wednesday)

Keynote , Sala/Room: A09:00-10:00 Deep Learning for Multimedia Data: Teaching Computers to Sense.

Omar Florez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

XXIII Simposio Iberoamericano de Educacion Superior en ComputacionSession: S1, Sala/Room: CChair: Tony Clear (New Zealand)

10:15-10:45 Philosophy of Computer Science and Its Impact on Education: Towardsthe construction of an interdisciplinary team. Sylvia da Rosa, Federico Go-mez, Alejandro Chmiel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

10:45-11:15 Adoption alternatives of academic innovations in computer scienceschools in Peru. Jorge Alvarado . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40

11:15-11:45 Educational Web Tool for Digital Image Processing. Martin Poletti, Hora-cio Legal, Jacques Facon, Claudio Barua, Jose Luis Vazquez . . . . . . . . . . . . . . . . . . . . 40


10:15-10:45 Forecast flows in a section of the Bogota River by Artificial IntelligentSystems. William Moscoso, Luis Mauricio Agudelo . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

10:45-11:15 An Application of ILS heuristic to Periodic Vehicle Routing Problemwith Heterogeneous Fleet and Fixed Costs. Robert Cristian Abreu, Jose EliasClaudio Arroyo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

11:15-11:45 A Variable Neighborhood Search Heuristic for the Traveling SalesmanProblem with Hotel Selection. Marques Sousa, Luiz Satoru Ochi, Igor Macha-do Coelho, Luciana Brugiolo Goncalves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Simposio Latinoamericano de Manejo de Datos e InformacionSession: S1, Sala/Room: BChair: Renzo Angles (Chile)

10:15-10:45 Semantic Recommender System for the Recovery of the Preserved WebHeritage. Jose Aguilar, Claudia Leon, Omar Portilla . . . . . . . . . . . . . . . . . . . . . . . . . . 36

10:45-11:15 Distributed Directory System: A Healthcare Use Case for Rural Areas.Alethia Hume, Fausto Giunchiglia, Luca Cernuzzi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36

11:15-11:45 MineraSkills: Mineracao de Dados Aplicada as Vagas Anunciadas noLinkedIn Visando Definir o Perfil Profissional. Dayane Cristine Caldeira,Ronaldo Correia, Rogerio Eduardo Garcia, Danilo Eler, Celso Olivete . . . . . . . . . . . 36

11:45-12:15 A Proposal for Customizing Queries on XML documents based on Con-ditional Preferences. Angelica Medeiros, Valeria Soares, Eudisley Anjos . . . . . . 36

Collaboration Networks in ComputingSession: S1, Sala/Room: DChair: Adenilso Simao (Brazil)

xii

10:15-10:45 USP Brasil: Redes de Colaboracao na Area de Computacao: perspecti-vas e acoes. Adenilso Simao . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

10:45-11:15 UChile Chile: Programa de Doctorado en Ciencias de la Computacionen la Universidad de Chile . Gonzalo Navarro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48

Keynote , Sala/Room: A14:00-15:00 Compact Data Structures. Gonzalo Navarro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3

Simposio Latinoamericano de Computacion Grafica, Realidad Virtual y Procesamientode ImagenesSession: S1, Sala/Room: DChair: Alex Cuadros-Vargas (Peru)

15:15-15:45 Indexacion de Imagenes Faciales mediante Algoritmo basado en Permu-taciones. Christian von Lucken, Liz Gonzalez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

15:45-16:15 A heuristic model for determining the sperm motility grade by videotracking. Diego Garate, Rosario Medina Rodrıguez, Filomen Incahuanaco, CesarBeltran . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

16:15-16:45 Automatic Detection of Glaucoma Using Disc Optic Segmentation andFeature Extraction. Maila Claro, L. Santos, Wallinson Silva, Flavio Henrique,Alcilene Sousa, Andre Macedo Santana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

16:45-17:15 3D Mesh Interest Point Detection using GISIFs and Heat Diffusion. JanJose Hurtado Jauregui, Cristian Lopez del Alamo, Madeley Karina Coaquira Con-gona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

Tutorial , Sala/Room: A15:15-19:15 Seguridad en la Web. Ernst Leiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Simposio Latinoamericano de Manejo de Datos e InformacionSession: S2, Sala/Room: BChair: Renzo Angles (Chile)

15:15-15:45 Towards Semantic Social Networks. Ronald Chenu-Abente, Fausto Giunchi-glia, Luca Cernuzzi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

15:45-16:15 Prediction of Tourist Traffic to Peru by using Sentiment Analysis inTwitter Social Network. Ricardo Linares, Jose Herrera, Ana Cuadros, Luis Al-faro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

16:15-16:45 Revisiting the Visibility Problem with a Hybrid Structure Paradigm.Icaro da Cunha, Luiz Goncalves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

XXIII Simposio Iberoamericano de Educacion Superior en ComputacionSession: S2, Sala/Room: CChair: Ernesto Cuadros-Vargas (Peru)

15:15-15:45 Student Understanding of the C++ Notional Machine Through Tradi-tional Teaching with Conceptual Contraposition and Program MemoryTracing. Jeisson Hidalgo-Cespedes, Gabriela Marın, Vladimir Lara . . . . . . . . . . . . . 40

15:45-16:15 Academic Performance of University Students and its Relation withEmployment . Laura Lanzarini, Marıa Emilia Charnelli, Javier Dıaz . . . . . . . . . . 41

16:15-16:45 University Bonding with Productive Sector Companies: A literature re-view. Sandra Cabrera Alzate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Simposio Latinoamericano de Sistemas de Informacion de Gran EscalaSession: Tecnologıas, Sala/Room: BChair: Andrea Delgado (Uruguay)

17:15-17:45 CARMiCLOC: Context Awareness Middleware in CLOud Computing.Jose Aguilar, Marxjhony Jerez, Ernesto Exposito, Thierry Viellemur . . . . . . . . . . . . 27

17:45-18:15 Comparando ZeroMQ y RabbitMQ como tecnologıas orientadas a even-tos. Nicolas Estrada, Hernan Astudillo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

18:15-18:45 Case-based Reasoning for Web Service Discovery and Selection. Alan DeRenzis, Martin Garriga, Andres Flores, Alejandra Cechich, Alejandro Zunino . . . 27

V Workshop en Nomenclatura y Acreditacion en Programas de Computacion

xiii

Session: S2, Sala/Room: CChair: Ernesto Cuadros-Vargas (Peru)

17:15-19:15 Discusion y resultados. Ernesto Cuadros-Vargas, Ariel Sabiguero Yawelak . . . .42

October 22st (Jueves/Quinta feira/Thursday)

Keynote , Sala/Room: A09:00-10:00 Combining Matching Dependencies and Machine Learning via Datalog

for Entity Resolution in Databases. Leopoldo Bertossi . . . . . . . . . . . . . . . . . . . . . . . 4

Premios de Investigacion de Google para America LatinaSession: Google Symposium, Sala/Room: CChair: Susana Pabon (Google)

10:15-10:45 Learning Dynamic Action Units for Three-dimensional Facial ExpressionRecognition. Pablo Arbelaez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49

10:45-11:15 Gradual Security Typing for the Web. Raimil Cruz Concepcion . . . . . . . . . . . .4911:15-11:45 Interconnected Dual Biosensor for Type II Diabetes Mellitus. Mathieu

Hautefeuille . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4911:45-12:15 Urban Coordination of Autonomous Vehicles. Jorge Zapotecatl . . . . . . . . . . . 49

I Concurso Latinoamericano de Tesis de DoctoradoSession: S1, Sala/Room: AChair: Rodrigo Santos (Argentina)

10:15-10:45 Multivariate Investigation of NP-Hard Problems: Boundaries BetweenParameterized Tractability and Intractability. Ueverton dos Santos Souza 46

10:45-11:15 Dynamic Composition of REST services. Jesus Bellido . . . . . . . . . . . . . . . . . . . . 4611:15-11:45 Graph Laplacian for Spectral Clustering and Seeded Image Segmenta-

tion. Wallace Casaca . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .46

Simposio Latinoamericano de Sistemas de Informacion de Gran EscalaSession: Aplicaciones, Sala/Room: BChair: Hernan Astudillo (Chile)

10:15-10:45 Monitoring and Enforcing Data Protection Laws within an E-government Interoperabilty Platform. Andres Echevarrıa, Dahiana Morales,Laura Gonzalez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

10:45-11:15 Chilean Virtual Observatory. Mauricio Solar, Mauricio Araya, Luis Arevalo,Ricardo Contreras, Victor Parada, Diego Mardones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

11:15-11:45 Business Intelligence applied to Learning Analytics in student-centeredlearning processes.. Guido Riofrio, Angel Encalada, Daniel Guaman, Jose Agui-lar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28


10:15-10:45 Deteccao Automatica do Glaucoma pelo uso de Descritores Locais, Ma-triz GLCM e Aprendizado de Maquina. Wallinson Silva, Maila Claro, L.Santos, Flavio Henrique, Alcilene Sousa, Nayara Moura . . . . . . . . . . . . . . . . . . . . . . . . . .9

10:45-11:15 A Texture and Curvature Bimodal Leaf Recognition Model for Identi-fication of Costa Rican Plant Species. Erick Mata-Montero, Jose Carranza-Rojas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

11:15-11:45 An Open Source Framework To Manage Kinect On The Web. FranciscoMoreno, Esmitt Ramırez, Francisco Sans, Rhadames Carmona . . . . . . . . . . . . . . . . . . . 9

Keynote , Sala/Room: A14:00-15:00 Learning to select learning algorithms. Andre de Carvalho . . . . . . . . . . . . . . . . . . 4

XXII Concurso Latinoamericano de Tesis de MaestrıaSession: S1, Sala/Room: A

xiv

Chair: Andre de Carvalho (Brazil)15:15-15:45 Espelho virtual interativo para simulacao de maquiagem. Filipe Morgado

Simoes de Campos, Carlos H. Morimoto . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4515:45-16:15 Image Segmentation by Image Foresting Transform with Non-smooth

Connectivity Functions. Lucy A. Choque Mansilla . . . . . . . . . . . . . . . . . . . . . . . . . . . 4516:15-16:45 Solving the Art Gallery Problem: A Practical and Robust Method for

Optimal Point Guard Positioning. Davi Colli Tozoni . . . . . . . . . . . . . . . . . . . . . . . 45


15:15-15:45 An Efficient approach for interest points detection in non-rigid shapes.Cristian Lopez del Alamo, Luciano Arnaldo Romero Calla, Lizeth Joseline FuentesPerez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

15:45-16:15 Computerized Diagnosis of Melanocytic Lesions Based on the ABCDMethod. Deysi Leguizamon, Laura Bareiro, Jose Luis Vazquez Noguera, Diego P.Pinto-Roa, Liza A. Salgueiro Toledo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

16:15-16:45 Speeding up the combination of multiple descriptors for different boun-dary conditions. Rodrigo Barroso, Marcelo Ponciano-Silva, Agma Traina, RenatoBueno . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Simposio Latinoamericano de Sistemas de Informacion de Gran EscalaSession: BPM, Sala/Room: BChair: Jose Aguilar (Venezuela)

15:15-15:45 Automating the process of building flexible Web Warehouses with BPMSystems. Andrea Delgado, Adriana Marotta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28

15:45-16:15 Hacia una ontologıa de arquitectura de negocio y de procesos basada enun modelo de negocio. Bernhard Hitpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29

Keynote , Sala/Room: A17:15-18:15 XML Semantic Disambiguation: Background, Applications, and Ongoing

Challenges. Richard Chbeir . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

October 23th (Viernes/Sexta feira/Friday)

Keynote , Sala/Room: A09:00-10:00 Testing Based on Finite State Machines: Past, Present and Future. Ade-

nilso Simao . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Tutorial , Sala/Room: B10:15-11:15 Digital Disruption. Juan Pablo Guizado Serrano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Keynote , Sala/Room: A10:15-11:15 Cloud Security and Forensics. George J. Proeller . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

xv

List of Authors (332)Abarca, Melissa: p.23Abelem, Antonio: p.34Abreu, Robert Cristian: p.20Aceval, Cristian: p.11Agudelo, Luis Mauricio: p.19Aguilar, Jose: p.18, p.36, p.28, p.27Alfaro, Luis: p.37Alfaro-Velasco, Jorge: p.32Alferillo, Viviana: p.30Altamiranda Perez, Junior Amilcar: p.18Alvarado, Jorge: p.40Alves, Tiago: p.21Anjos, Eudisley: p.36Araya, Mauricio: p.28Arbelaez, Pablo: p.49Arboleda, Hugo: p.38Arevalo, Luis: p.28Arevalos, Sara: p.13Aroca Acosta, Richard: p.42Arroyo, Jose Elias Claudio: p.20Arteta, Adolfo: p.12Astudillo, Hernan: p.27Avila-Herrera, Juan: p.17Ayala, Victor: p.13Ayala-Rincon, Mauricio: p.38

Balke, Maicon: p.30Ballesteros Ricaurte, Javier Antonio: p.21Bareiro, Laura: p.10Barroso, Rodrigo: p.37Barua, Claudio: p.40Baran, Benjamın: p.16, p.18, p.13, p.11, p.15Baran, Benjamın: p.13Bautista Ubillus, Efraın: p.31Bazan, Patricia: p.34Bellido, Jesus: p.46Beltran, Cesar: p.8Belzarena, Pablo: p.12Benner, Peter: p.14Bermeo, Jorge L.: p.21Bernal, Ruben: p.19Bertossi, Leopoldo: p.4Bogaz Zarpelao, Bruno: p.13Brandao, Marcius: p.31Brugiolo Goncalves, Luciana: p.20Bruschi, Sarita M.: p.32Bueno, Renato: p.37Baez Gonzalez, Maximiliano: p.24

Cabrera Alzate, Sandra: p.41Caldeira, Dayane Cristine: p.36Campos, Gustavo: p.18Campuzano, Alfredo: p.23Cancela, Hector: p.16

Canhadas Genvigir, Elias: p.30Cano, Christian: p.31Cappo, Cristian: p.14Cardenas, Andrea: p.22Cardoso, Ismael: p.21Carmona, Rhadames: p.9Carranza-Rojas, Jose: p.9Casaca, Wallace: p.46Casas, David: p.38Cechich, Alejandra: p.27Cernuzzi, Luca: p.36, p.37Charnelli, Marıa Emilia: p.41Chbeir, Richard: p.4Chenu-Abente, Ronald: p.37Cherman, Everton: p.16Chmiel, Alejandro: p.40Choque Mansilla, Lucy A.: p.45Claro, Maila: p.8, p.9Clua, Esteban: p.1Coaquira Congona, Madeley Karina: p.9Condori-Fernandez, Nelly: p.2Contreras, Ricardo: p.28Correia, Ronaldo: p.36Cortes, Mariela: p.18, p.31Costa, Heitor: p.33, p.32Cruz, Rene: p.15Cruz Concepcion, Raimil: p.49Cruz Quispe, Laura: p.19Cuadros, Ana: p.37Cuadros-Vargas, Ernesto: p.42, p.42, p.42, p.42Cuba, Jaime: p.15Cunha de Miranda, Leonardo: p.25, p.25, p.25

da Costa, Cristiano Andre: p.21, p.11da Costa Vieira Rezende, Josiane: p.16da Cunha, Icaro: p.8da Rosa, Sylvia: p.40da Silva, Anderson Faustino: p.14da Silveira, Lucas: p.38de Campos, Filipe Morgado Simoes: p.45de Carvalho, Andre: p.4de Carvalho Jr, Antonio: p.24De la Cruz, Danni: p.22de Lima, Thaynara A.: p.38de Lima, Ewerton Daniel: p.14de Menezes, Bruna: p.25, p.25, p.25De Renzis, Alan: p.27Delgado, Andrea: p.28, p.44Diaz, Javier: p.34Dornelas, Jairo Simiao: p.33dos Santos, Andre Gustavo: p.17, p.18dos Santos Souza, Ueverton: p.46dos Santos Viana, Renan Jose: p.18

xvi

Dourado, George G. M.: p.32Dufrechou, Ernesto: p.14Davalos Gimenez, Enrique: p.11, p.15Davila, Abraham: p.31, p.34Dıaz, Javier: p.41

Echevarrıa, Andres: p.27Egidio Cunha, Luiz: p.26Ehlers, Ricardo: p.11Eler, Danilo: p.36Encalada, Angel: p.28Esteves Cunha de Miranda, Erica: p.25Esteves Cunha de Miranda, Erica: p.25, p.25Estrada, Nicolas: p.27Estrella, Julio: p.32, p.11Exposito, Ernesto: p.27

Fabri, Jose Augusto: p.30Facon, Jacques: p.40Faustino da Silva, Anderson: p.13Fernandez, Romina: p.24Fernandez Fabian, Karla Mariel: p.23Filippi, Lucas: p.34Flores, Andres: p.27Flores, Pablo: p.12Florez, Omar: p.2Fontoura, Lisandra: p.30Franco, Victor: p.11Freire, Emmanuel Savio Silva: p.18, p.31Freitas Souza, Marcone: p.16Fuentes Perez, Lizeth Joseline: p.10

Garcia, Rogerio Eduardo: p.36Garriga, Martin: p.27Giandini, Roxana: p.34Gimenez, Jose: p.14Giunchiglia, Fausto: p.36, p.37Gonzalez, Liz: p.8Gonzalez, Laura: p.27Gonzalez Rodas, Guillermo: p.24Gonzalez-Barbone, Vıctor: p.12Goncalves, Ramon: p.32Goncalves, Luiz: p.8Groterhorst, Vitor: p.23Guaman, Daniel: p.28Guimaraes, Marcelo: p.26Guizado Serrano, Juan Pablo: p.7Garate, Diego: p.8Gomez, Federico: p.40Gomez, Gabriel: p.12

Hautefeuille, Mathieu: p.49Henrique, Flavio: p.8, p.9Hernandez, Alejandro: p.22Herrera, Jose: p.23, p.37Hidalgo-Cespedes, Jeisson: p.40Hideyuki Yokoyama, Marcos: p.26

Hitpass, Bernhard: p.29Huaynacho, Yoni: p.15Hume, Alethia: p.36Hurtado Jauregui, Jan Jose: p.9

Ilıdio da Silva, Rone: p.16Incahuanaco, Filomen: p.8

Jenkins, Marcelo: p.32Jerez, Marxjhony: p.27Junior, Robert: p.18

Karanik, Marcelo: p.19Kemper, Nicolas: p.2

L’Erario, Alexandre: p.30La Serna Palomino, Nora: p.17, p.31Lanzarini, Laura: p.41Lara, Vladimir: p.40Larroca, Federico: p.12Legal, Horacio: p.40Leguizamon, Deysi: p.10Leiss, Ernst: p.6Leon, Claudia: p.36Lima, Joaquın: p.23Lima, Igor: p.32Linares, Ricardo: p.37Lopes, Charlie Silva: p.33Lopes de Souza, Paulo S. L.: p.32Lopez-Pires, Fabio: p.13Lourenco, Joao: p.32Lucio, Diego Rafael: p.19Lund, Maria Ines: p.30Lopez, Ruben: p.23Lopez del Alamo, Cristian: p.9, p.10

Macedo Santana, Andre: p.8Machado Coelho, Igor: p.20Maguina Perez, Rolando Alberto: p.17Maldonado, Jorge J.: p.21Maldonado e Gomes da Costa, Yandre: p.19Mals, Nadav: p.23Mamani-Aliaga, Alvaro: p.15Mardones, Diego: p.28Marotta, Adriana: p.28Martinez Garro, Jose: p.34Martins, Valeria: p.26Martınez Porras, Alexandra: p.32Marın, Gabriela: p.40, p.43Mata, Francisco J.: p.22Mata-Montero, Erick: p.9Matturro, Gerardo: p.30Mazon, Jose Norberto: p.23Medeiros, Angelica: p.36Medina Rodrıguez, Rosario: p.8Mejıa, Paul: p.22Mejıa, Magali: p.21

xvii

Melgar, Andres: p.31Menchaca-Mendez, Rolando: p.12Mendes, Nilson: p.17Messias, Valter: p.11Mine Subasi, Munevver: p.17Miranda de Barros, Rodolfo: p.13Monard, Maria Carolina: p.16, p.43Monteiro, Ana Maria: p.43Montoya, J. Andres: p.38, p.38Morales, Dahiana: p.27Moreno, Francisco: p.9Morimoto, Carlos H.: p.45Moscoso, William: p.19Mota, Bruno: p.21Moura, Nayara: p.9Moura, Raimundo: p.30Murray, Leslie: p.16Mendez-Porras, Abel: p.32

Navarro, Gonzalo: p.3, p.48Nina, Wilder: p.15Nogueira Luz, Bruno: p.26Nunes Marciano, Juvane: p.25, p.25, p.25Nunez, Tania: p.13

Ochoa Luna, Jose Eduardo: p.19Oliveira, Jaime: p.25Olivete, Celso: p.36Oviedo, Edgar: p.23

Paciello, Julio: p.13Parada, Victor: p.28Paredes, Nancy: p.22Patino, Raquel: p.44Paz, Andres: p.38Pelaez Sanchez, Jose Ignacio: p.19Pineda-Briseno, Anabel: p.12Pinto, Francisco Chaves: p.33Pinto-Roa, Diego P.: p.12, p.10Poletti, Martin: p.40Ponciano-Silva, Marcelo: p.37Portilla, Omar: p.36Prado, Rafael R.: p.32Proeller, George J.: p.5Padua, Wilson: p.31

Queiroz Junior, Nilton: p.13Quesada, Ariella: p.22Quintana, Enrique: p.14Quirino, Guilherme: p.23

Ramos Isique, Luis Horacio: p.13Ramırez, Esmitt: p.9Rangel, Carlos: p.18Reis, Rodrigo: p.34Remon, Alfredo: p.14Riano Herrera, Jairo Armando: p.21

Ribeiro, Moshe: p.34Righi, Rodrigo da Rosa: p.21, p.21, p.11Rigo, Sandro Jose: p.34Riofrio, Guido: p.28Riquelme, Nery: p.18Rocha Marques, Fabio: p.25Rocha Marques, Fabio: p.25Rodrigues, Vinicius Facco: p.11Rodrıguez, Guillermo: p.22Romero Calla, Luciano Arnaldo: p.10Rostirolla, Gustavo: p.11Rubino, Gerardo: p.16Rubira, Cecılia: p.33Re, Reginaldo: p.32

Sabiguero Yawelak, Ariel: p.42, p.42, p.42, p.42Salgueiro Toledo, Liza A.: p.10Sans, Francisco: p.9Santos, Rafael: p.26Santos, Rodrigo: p.6Santos, L.: p.8, p.9Sartorio, Alejandro: p.22Satin, Ricardo F. P.: p.32Satoru Ochi, Luiz: p.20Schneck de Paula Pessoa, Marcelo: p.31, p.34Segovia, Edison: p.22Segundo, Hildebrando: p.11Serrano, Juber: p.15Silva, Wallinson: p.8, p.9Silva, Raul: p.33Silva, Ivan Saraiva: p.11Silva, Juan Pablo: p.14Silva, Luis Alvaro de Lima: p.30Silva Areao, Andreza: p.26Silveira, Denis Silva da: p.33Silveira Duarte, Alessandro: p.30Simao, Adenilso: p.5, p.48Soares, Samuel A.: p.31Soares, Helcio A.: p.30Soares, Valeria: p.36Sobrevilla Cabezudo, Marco: p.17Solar, Mauricio: p.28Soncco-Alvarez, Jose Luis: p.38Sousa, Alcilene: p.8, p.9Sousa, Marques: p.20Souza, Simone R. S.: p.32Souza, Luiz Sergio de: p.23Souza, Solange N. Alves de: p.23

Tedini, Daniel: p.22Tilerıa, Marcos: p.15Tirado, Francisco: p.1Tobar, Johanna: p.22Torres Bobadilla, Mateo: p.16Tozoni, Davi Colli: p.45Traina, Agma: p.37

xviii

Tsoumakas, Grigorios: p.16

Vaquero, Marcelo: p.22Victoria Barbosa, Jorge Luis: p.21, p.21, p.11Viellemur, Thierry: p.27von Lucken, Christian: p.24von Lucken, Christian: p.18, p.8Vazquez Noguera, Jose Luis: p.40, p.10

Wiese, Igor Scaliante: p.32

Yakaryilmaz, Abuzer: p.38Yanez Benavides, Estela del Rocio: p.19Yanyachi, Pablo: p.15Yari, Yessenia: p.15Yu, Aloysius: p.15

Zapotecatl, Jorge: p.49Zela Ruiz, Jael: p.33Zubcoff, Jose Jacobo: p.23Zunino, Alejandro: p.27

Keynote

Keynote

Next Generation of GPUsEsteban Clua1

1Federal Fluminense University. Rio de JaneiroBrazil

email: [email protected]

Schedule:Mon 19th@09:00, Room: A

GPUs were initially developed for solving real ti-me graphics pipeline. Due their massively parallelarchitectures and the creation of a unified archi-tecture, they became an important alternative forhigh performance computing. Modern GPUs ha-ve more than 3000 cores and may achieve up to7TFlops, which is dozens of times the amount ofa CPU based architecture. This talk will brieflyshow the hardware architecture of a modern GPU,will discuss about trends and the future of thenext generation of GPUs and will present howdifferent areas are being enhanced by this tech-nologies, such as video-games, deep-learning andself-driving cars.

Short Biography

Esteban Walter Gonzalez Clua is graduated inComputer Science at Universidade de Sao Pau-lo and has master’s and PhD degree in Compu-ter Science. Has experience in Computer Scienceand has being acting in the following subjects: ga-me engine architectures, 3D games, GPU Compu-ting, education with games, real time computergraphics and digital entertainment. Today Este-ban is associated professor at the computer scien-ce of Universidade Federal Fluminense, in Rio deJaneiro, and director of UFF Medialab. Estebanis one of the founders of SBGames - BrazilianSymposium of Digital Entertainment and VideoGames (the largest conference in the subject inSouth America), president of the Brazilian Com-puting Society Game Committee and member ofprogram committees of many conferences in VideoGames. In 2007 received the prize of the persona-lity which most contributed for the growth of thevideo game industry in Brazil and in 2009 and2013 received the prize of Young Scientist of theState of Rio de Janeiro. Esteban is the coordinatorof the first Latin America NVIdia Center of Exce-lence, which is in UFF Medialab and is a NVIDIAFellow professor.

Los procesadores en el final de la ley deMoore

Francisco Tirado11Universidad Complutense. Madrid Spain



Desde su aparicion, los procesadores han dobladosu velocidad cada 18 meses. Esto ha sido debi-do, por una parte, a los avances en la integracionde circuitos que permiten duplicar el numero detransistores cada 18 meses y una mayor velocidaddel reloj y por otra, a mejoras de su arquitectura( organizacion interna de los diferentes modulosoperativos) orientadas a ejecutar un numero cadavez mayor de instrucciones por ciclo. La dedica-cion de recursos para incrementar el rendimientode un procesador ha llegado a un punto donde elretorno que se obtiene en rendimiento es pequeno.Hoy en dıa, factores como la potencia y disipaciontermica son factores determinantes en el disenode los chips. La explotacion, por tanto de estosrecursos adicionales se ha orientado a la imple-mentacion en el chip de multiples niveles de pa-ralelismo. Ademas, nos encontramos en el posiblefin de la Ley de Moore, y esto plantea problemasadicionales en el escalado tecnologico, dejandosede cumplir la reglas que ha funcionado por masde 50 anos

Short Biography

Francisco Tirado es Catedratico de Arquitecturay Tecnologıa de Computadores en la UniversidadComplutense de Madrid desde 1986. Ha trabaja-do en diferentes areas de investigacion dentro dela arquitectura de sistemas, computacion de altasprestaciones y arquitectura del procesador. Hastala actualidad es coautor de mas de 250 publicacio-nes en revistas y conferencias internacionales. Haparticipado en la organizacion de mas de 90 Con-gresos Internacionales como General Chair, Pro-gram Chair, miembro del Comite de Programa,Chair de Sesion, Conferenciante Invitado y revi-sor de artıculos. Ha impartido mas 80 conferen-cias en congresos internacionales y universidades.Es Premio Nacional de Informatica 2013 y DoctorHonoris Causa por la Universidades de San Agus-tın ( Peru) y la Nacional de Asuncion ( Paraguay). Ha ocupado diferentes puestos en organismos deinvestigacion y evaluacion. Gestor en el ProgramaNacional de I+D, Presidente de la Comision Na-cional de Evaluacion de Actividad Investigadora,Director del Parque Cientıfico de Madrid. En laactualidad es Vicerrector de Investigacion de laUCM, Presidente de Honor de la Sociedad Cien-tıfica Informatica de Espana (SCIE) y miembrode la Comision Ejecutiva de la Confederacion deSociedades Cientıficas de Espana ( COSCE)

1

http://lattes.cnpq.br/4791589931798048

https://artecs.dacya.ucm.es/

www.ucm.esl

Keynote

Service Quality Assurance inContext-Aware SystemsNelly Condori-Fernandez1

1VU University Amsterdam. The Netherlandsemail: [email protected]

Schedule:Tue 20th@09:00, Room: A

Context-aware computing and associated techno-logies are becoming part of daily life. In fact,context-aware computing is considered to be agame-changing opportunity for enterprises to im-prove both productivity and profits. However, thispositive economic growth will also bring a num-ber of issues related to social sustainability (e.g.perceived enjoyment, resilience, privacy).

Given that evidence has corroborated that fee-lings and emotions dictate to a large extent our ac-tions and decisions, sustainability of context awa-re systems is approached from the perspective ofone of its ingredients, namely social acceptance. Inthis talk, I will introduce a sustainability assuran-ce framework that exploit emotional informationto adjust service quality levels at runtime.

Inteligencia Artificial en el Siglo XXINicolas Kemper1

1Universidad Nacional Autonoma de Mexico.Mexico Mexico



En esta conferencia se trata de proporcionar unaperspectiva apropiada y moderna de la inteligen-cia artificial (IA). Revisamos lo que hemos sidoy lo que se ha logrado, se comentan los logrosde la IA en el medio siglo pasado en lo que serefiere a busqueda, representacion de conocimien-to, razonamiento, aprendizaje, planeacion, etc. Sehace un recuento de la IA considerando el para-digma inicial como Inteligencia Artificial Simboli-ca, pasando por lo que se denomina InteligenciaComputacional y se culmina con el enfoque de In-teligencia Artificial Bioinspirada. Tambien sope-samos nuestras perspectivas actuales bajo la metade lograr el nivel de una IA mas humana. En to-do este panorama reflexionamos si la creacion deun nivel humano de IA es imposible, o se darancreaciones exitosas en un futuro muy cercano.

Short Biography

Ingeniero Industrial, Universidad Nacional deTrujillo de Peru, Maestrıa en Ingenierıa, espe-cialidad sistemas expertos aplicados en la indus-tria electrica y Doctorado en Ingenierıa, especia-lidad en Inteligencia Artificial aplicada en siste-mas energeticos, ambos grados en la Universidad

Nacional Autonoma de Mexico. Sus lıneas de in-vestigacion y desarrollo tecnologico son: Desarro-llo de sistemas inteligentes aplicados. Gestion delConocimiento y Organizaciones Inteligentes. Ma-nufactura Inteligente, Instrumentacion Virtual In-teligente y Robotica Adaptativa. Minerıa de datose Inteligencia de Negocios. Dispositivos de vuelono tripulados autonomos para monitoreo espacial.Jefe del Departamento de Tecnologıas de la In-formacion y del Grupo de Sistemas Inteligentesde este Centro de Investigacion. Profesor y Tu-tor del Postgrado en Ciencias e Ingenierıa de laComputacion en la misma UNAM. Tiene alrede-dor de 30 artıculos en memorias de eventos y re-vistas internacionales y regionales, alrededor de 60reportes tecnicos de proyectos de desarrollo tec-nologico, ha desarrollado diversos prototipos tec-nologicos y software inteligente. Ha publicado ellibro “Sistema Inteligente para la Gestion de unaPlanta Geotermoeletrica”, y esta en proceso de pu-blicacion los libros “Manual para el Desarrollo deSistemas Expertos”, “Desarrollo de Sistemas Difu-sos”, “Inteligencia Artificial Aplicada”, “Desarrollode Proyectos de Minerıa de Datos” y “Evaluacionde Proyectos Energeticos”. Miembro del Techni-cal Committee on Artificial Intelligence and Ex-pert System, periodo 2000-2005, de la Internatio-nal Association of Science and Technology for De-velopment (IASTED) de Canada. Miembro del In-ternational Program Committee (IPC) de diversoseventos internacionales. Miembro de la AsociacionMexicana de Economıa en Energıa; del Interna-tional Association of Science and Technology forDevelopment (IASTED), Canada; Miembro Fun-dador de la Academia Mexicana de Tecnologıa;Miembro del Institute for Systems and Technolo-gies of Information, Control and Communicationde Portugal (INSTICC) y es Miembro Fundadorde la Sociedad Peruana de Inteligencia Artificial.Ha tenido diversas distinciones como la MedallaGabino Barreda por el mejor promedio en sus es-tudios de maestrıa en Ingenierıa. Ası fue honradocon el grado de Doctor Honoris Causa por la Uni-versidad Nacional de Piura, Peru.

Deep Learning for Multimedia Data:Teaching Computers to Sense

Omar Florez11Intel Labs. California USA


Schedule:Wed 21st@09:00, Room: A

For the past few years, deep learning has been ma-king rapid progress in both techniques and appli-cations; significant performance gains were repor-ted using deep learning in automatic speech recog-

2

http://www.ccadet.unam.mx/secciones/depar/sub4/sinte/sint.html

www.unam.mx

www.intel.com

Keynote

nition and image recognition over hand-optimizedfeature representations.

Advances in smartphones, tablets, and weara-bles have made possible to sense a rich collectionof user data, examples include audio, images, andvideo. This information allows us to infer user con-texts such as faces, semantic locations, activities,and mood states enabling better and personalizeduser experiences. This explains the growing inter-est of industry (Intel, Facebook, Google, NVIDIA,Spotify, Netflix, Baidu, etc.) trying to take advan-tage of deep learning capabilities in recent years.

In several image and speech tasks, the successof deep learning is due to its ability to learn re-presentations from noisy and unstructured data.Context sensing faces the same problem therefo-re we believe applications of deep learning in thisdomain can be advantageous. During this talk wewill try bringing together researchers and appli-cants to discuss some of the deep learning algo-rithms and capabilities for multimedia and con-text domains as well as explore possible new re-search areas.

Short Biography

Dr. Omar U. Florez is a Research Scientist atthe Anticipatory Computing Group at Intel Labs(California, USA). He graduated from Universi-dad Nacional de San Agustin, Peru in 2007 andreceived his Ph.D. in Computer Science at UtahState University in 2013. He is a recipient of an In-novation Award on Large-Scale Analytics by IBMResearch, and the organizer of the NSF-fundedBroader Participation in Data Mining workshopat KDD in 2014, which for first time funded the at-tendance of under-represented researchers world-wide. He is also the co-founder of South Ameri-cans in Computing. Dr. Florez’s research interestscover statistical machine learning, recommendersystems, and deep learning for multimedia data.He has 20+ academic publications and journals inACM, IEEE, and Springer.

Compact Data StructuresGonzalo Navarro1

1Universidad de Chile. Santiago de Chile Chileemail: [email protected]


Compact Data Structures is a new and excitingresearch area that lies in the intersection of Algo-rithms and Data Structures, Information Theory,and Data Compression. It has gained much atten-tion in recent years due to the sharp increase inthe sizes of the datasets available to applications

and the widening gap in the memory hierarchy,which makes it much faster to process data inmain memory than in external storage. CompactData Structures seek for compressed data repre-sentations that can be manipulated and querieddirectly in compressed form, therefore enablingthe representation of much larger datasets in mainmemory. I will describe the main ideas of the areaand the two most relevant success stories: the re-presentation of n-node trees in 2n bits and therepresentation of powerful text indexes within thespace of the compressed text.

Short Biography

Gonzalo Navarro completed his PhD in ComputerScience in 1998 at the University of Chile, wherehe is currently full professor. His areas of interestinclude algorithms and data structures, text sear-ching, compression, and metric space searching.He has directed the Millennium Nucleus Centerfor Web Research, RIBIDI (an Ibero Americanproject funded by CYTED), and a project fun-ded by Yahoo! Research, apart from smaller pro-jects. He has participated in various research pro-jects, such as the Millennium Institute for Cell Dy-namics and Biotechnology, an ECOS/CONICYTproject (Chile-France cooperation), AMYRI (aCYTED project), and a Fondef project. Currently,he participates in the Millennium Nucleus Infor-mation and Coordination in Networks and in theCenter for Biotechnology and Bioengineering.

He has been PC (co-)chair of several conferen-ces: SPIRE 2001, SCCC 2004, SPIRE 2005, SI-GIR 2005 Posters, IFIP TCS 2006, a track of ENC2007, SISAP 2008, SISAP 2012, and LATIN 2016.He co-created SISAP on 2008, and was SteeringCommittee member of SPIRE, LATIN, and SI-SAP. He is a member of the Editorial Board of thejournals Information Retrieval, ACM Journal ofExperimental Algorithmics, and Information Sys-tems. He has been guest editor of special issues inACM SIGSPATIAL and Journal of Discrete Algo-rithmics. He has been PC member of more than 50international conferences and reviewer for about40 international journals. He has given around 50invited talks in several universities and interna-tional conferences, including 10 plenary talks and3 tutorials in international conferences. He crea-ted in 2005 the Workshop on Compression, Text,and Algorithms, which has become a permanentsatellite of SPIRE.

He has coauthored a book published by Cam-bridge University Press, about 20 book chapters, 7proceedings of international conferences (editor),more than 130 papers in international journals,and more than 200 in international conferences.

3

http://www.dcc.uchile.cl/~gnavarro

http://www.dcc.uchile.cl

Keynote

He is one of the most prolific and highly citedauthors in Latin America.

Combining Matching Dependencies andMachine Learning via Datalog for Entity

Resolution in DatabasesLeopoldo Bertossi1

1Carleton University. Ottawa Canadaemail: [email protected]

Schedule:Thu 22st@09:00, Room: A

In this presentation we give an introduction to thecombination of matching dependencies for entityresolution and support-vector machines, a classifi-cation method in machine learning. We show thatthey can be used to address the problems of iden-tifying duplicate representations in data sourcesand merging them into single representations. Thelogical glue is provided by Datalog, a logical querylanguage for relational databases, in its incarna-tion as LogiQL, which has been developed by Lo-gicBlox. This presentation emphasizes the currenttrend of logically combining different techniquesfor addressing different interrelated problems indata management.

The presentation will be given in Spanish.

Short Biography

Leopoldo Bertossi has been Full Professor at theSchool of Computer Science, Carleton University(Ottawa, Canada) since 2001. He is also a FacultyFellow of the IBM Center for Advanced Studies(IBM Toronto Lab). He has been a professor atthe Department of Computer Science, PUC-Chile(until 2001); and also the President of the ChileanComputer Science Society (SCCC). His researchinterests include data management in general, da-tabase theory, business intelligence, data quality,semantic web data, logic-based knowledge repre-sentation, and machine learning. He obtained aPhD in Mathematics from the PUC-Chile in 1988.

Learning to select learning algorithmsAndre de Carvalho1

1University of Sao Paulo. Sao Carlos-SP Brazilemail: [email protected]


A large number of learning algorithms have beendeveloped in the last decades and they have beenapplied to several tasks in different application do-mains, with different performance levels. Accor-ding to empirical and theoretical results, no singlealgorithm can outperform the others in every task.Thus, when using learning algorithms to solve a

new task, we are faced with the question of whichalgorithm to use. Metalearning provides a generalframework for the selection of the most suitable al-gorithm for a new task. This talk will discuss howmetalearning can be used for algorithm selectionin different learning tasks.

Short Biography

Andre C. P. L. F. de Carvalho is Full Professorin the department of Computer Science, Univer-sity of Sao Paulo, Brazil. His main research in-terests are data mining, data science and machi-ne learning. Prof. Andre de Carvalho has morethan 300 peer reviewed publications, including 10best papers awards from conferences organized byACM, IEEE and SBC. He is a member of the In-ternational Association for Statistical Computing(IASC) Council and director of the Center of Ma-chine Learning in Data Analysis of the Universityof Sao Paulo.

XML Semantic Disambiguation:Background, Applications, and Ongoing

ChallengesRichard Chbeir1

1University of Pau. Anglet Franceemail: [email protected]


Since the last two decades, XML has gained mo-mentum as the standard for Web information ma-nagement and complex data representation. Also,collaboratively built semi-structured informationresources such as Wikipedia have become preva-lent on the Web and can be inherently encodedin XML. Yet most methods for processing semi-structured information in general and XML inparticular handle mainly the syntactic propertiesof the data, while ignoring the semantics invol-ved. To devise more intelligent applications, oneneeds to augment syntactic features with machine-readable semantic meaning. This can be achievedthrough the computational identification of themeaning of data in context, also known as automa-ted semantic analysis and disambiguation, whichis nowadays one of the main challenges at the co-re of the Semantic Web. XML semantic-analysisprocessing and disambiguation become crucial inan array of applications ranging over semantic-aware query rewriting, semantic document cluste-ring and classification, schema matching, as wellas blog analysis and event detection in social net-works and tweets. This talk provides a conciseand comprehensive review of the methods relatedto XML-based semi-structured semantic analysis

4

http://scs.carleton.ca/~bertossi

http://scs.carleton.ca

http://www.icmc.usp.br/~andre

http://www.icmc.usp.br

http://www.univ-pau.fr/~rchbeir

http://liuppa.univ-pau.fr/live/

Keynote

and disambiguation. It will be composed of fourparts. First, I will briefly cover traditional wordsense disambiguation methods for processing flattextual data. Second, I will describe and catego-rize disambiguation techniques developed and ex-tended to handle semi-structured and XML da-ta. Third, I will describe current and potentialapplication scenarios that can benefit from XMLsemantic analysis. Fourth, I will discuss ongoingchallenges and future directions.

Short Biography

Richard Chbeir received his PhD in ComputerScience from the University of INSA DE LYON-FRANCE in 2001 and then his Habilitation de-gree in 2010 from the University of Bourgogne.He is currently a Full Professor in the ComputerScience Department in IUT de Bayonne in An-glet -France. His current research interests are inthe areas of social networks, multimedia seman-tics, XML and RSS similarity, and digital ecosys-tems. Richard Chbeir has published in interna-tional journals, books, and conferences, and hasserved on the program committees of several in-ternational conferences. He is currently the Chairof the French Chapter ACM SIGAPP.

Testing Based on Finite State Machines:Past, Present and Future

Adenilso Simao11University of Sao Paulo. Sao Carlos-SP Brazil


Schedule:Fri 23th@09:00, Room: A

State machines are among the simplest behavioralmodels which can be used for system modelling.Nonetheless, they are very expressive and have along history in computing science. The first tes-ting techniques based on this kind of models arefrom 50ties. Since then, several contributions ha-ve been proposed, including some recent results.In this talk, we are going to introduce the mainconcepts of testing based on state machines, inspecial, finite state machines. The main genera-tion methods will be discussed. The state-of-the-art methods will be presented and future direc-tions on this research area will be pointed out.

Short Biography

Possui graduacao em Bacharelado em Ciencia daComputacao pela Universidade Estadual de Ma-ringa (1997), mestrado em Ciencias da Compu-tacao e Matematica Computacional pela Univer-sidade de Sao Paulo (2000) e doutorado em Cien-cias da Computacao e Matematica Computacional

pela Universidade de Sao Paulo (2004). Realizouestagio de pos-doutoramento no Centre de Recher-che Informatique de Montreal (2008-2010). Desde2004 e Professor da Universidade de Sao Paulo(atualmente, e Professor Associado Nıvel 2). Temexperiencia na area de Ciencia da Computacao,com enfase em Engenharia de Software, atuan-do principalmente nos seguintes temas: teste desoftware, metodos formais e linguagens de progra-macao.

Cloud Security and ForensicsGeorge J. Proeller1

1Colorado Technical University. ColoradoSprings USA


Schedule:Fri 23th@10:15, Room: A

Cloud Computing offers an opportunity to increa-se capability with a promise of decreasing overallIT costs. The Cloud is also increasingly coupledwith the ubiquity of mobile computing and itsadvantages. Both are now a part of everyday li-fe - Amazon, Google, Dropbox, iTunes Cloud,RealAudio Cloud, etc. However the ”Benefits.ofthe Cloud cause us to forfeit aspects of control,visibility, and tracking data origin, source, and at-tribution. Traditional Computer Forensics requi-res physical access to systems for processes suchas disk imaging something not readily availablewhen working in the Cloud. Further the physicallocation of data can be very difficult to determi-ne in a private cloud and a near impossible taskin a public cloud where the data may not evenreside in the same country as the user. This talkdiscusses the challenges and opportunities withinthe emerging areas of cloud and mobile forensics.

Short Biography

Thirty-plus years’ experience in Information Tech-nology Systems (ITS) focusing on life-cycle In-formation Assurance/Computer Network Securityfor the Enterprise. He holds multiple security cer-tifications including the Certified Information Sys-tems Security Professional (CISSP) credential, theCertified Information Security Manager, and theGIAC Security Leadership Certification and Cer-tified Information Forensics Investigator and is apioneer in the transition of information securityto academia. His awards include being named aDistinguished Fellow of the Information SystemsSecurity Association (ISSA) and to the ISSA Hallof Fame.

5

http://www.icmc.usp.br

http://profs.etsmtl.ca/aabran/English/index.html

www.coloradotech.edu

Tutorial

Tutorial

Planificacion Optima de SistemasMultinucleo conrestricciones de calidad de

servicio, recursos y energıaRodrigo Santos1

1Universidad Nacional del Sur - CONICET.Bahıa Blanca Argentina



Dos hechos fundamentales contribuyeron al desa-rrollo de los sistemas moviles. Por un lado el aba-ratamiento del hardware y la expansion de la me-moria disponible. Por el otro la evolucion de lasarquitecturas. De hecho practicamente cualquiermicrocontrolador, aun aquellos economicos tienencaracterısticas multicore. Si bien puede que la uni-dad de procesamiento sea unica, cuentan con dis-positivos perifericos para las comunicaciones, eltratamiento de senales de entrada/salida y el ma-nejo de tiempos como mınimo.

La consolidacion de las tecnologıas de comu-nicacion inalambrica permitio el establecimientode redes de sensores y actuadores que miden, eva-luan y controlan procesos a distancia. Estos pro-cesos abarcan areas tan diversas como la indus-tria de procesos continuos al monitoreo ambiental,las redes sociales, la administracion del sistema detransporte publico o la atencion temprana de in-cidentes urbanos o ambientales. En esta clase desistemas la validacion temporal de la informacionadministrada y la coherencia de los mismos sonaspectos centrales para el funcionamiento del sis-tema.

Un adecuado analisis del comportamiento tem-poral de esta clase de sistemas requiere de polıticasde planificacion que puedan garantizar el cumpli-miento de los requerimientos. Se suma al problemade la planificacion, las restricciones que la calidadde servicio requerida y el consumo de energıa im-ponen en la actualidad sobre los sistemas.

En el curso se presenta la teorıa de planifica-cion clasica, los mecanismos de verificacion y va-lidacion de las mismas, los criterios de optimali-dad para sistemas de computo y comunicaciones.El modelado adecuado de esta clase de problemaspermite transferir resultados de un area a otra.

Short Biography

Dr. Rodrigo Santos es Prof. del Departamento deIngenierıa Electrica y de Computadoras de la Uni-versidad Nacional del Sur e Investigador Adjunto

del CONICET. Su area de interes es la planifica-cion de sistemas de tiempo real, el analisis de re-querimientos, los sistemas colaborativos y los sis-temas embebidos. Fue presidente de CLEI entre2008 y 2012, miembro del IEEE. Autor de variosartıculos en revistas y congresos de la especiali-dad.

Seguridad en la WebErnst Leiss1

1University of Houston. Houston USAemail: [email protected]


Empezamos con metodos para la proteccion decontenido digital, desde criptografıa hasta marcasde agua digitales. Despues hablamos sobre con-trol de acceso, incluyendo contrasenas y metodosbiometricos. De ahı pasamos a crımenes que in-volucran o dependen de la web. Terminamos conalgunos comentarios sobre privacidad y los efectosde la tecnologıa en el mundo cotidiano.

Short Biography

Ernst Leiss obtuvo grados academicos en inge-nierıa (Dipl.-Ing., TU Viena, 1975), ciencia de lacomputacion (M. Math., U. Waterloo, 1974), ymatematica (Dr. techn., TU Viena, 1976). Traba-jo como post doc en la Universidad de Waterloo(1976/7) y como profesor investigador en la Uni-versidad de Chile en Santiago (1978). Empezo enel departamento de ciencias de la computacion en1979 donde todavıa trabaja. Sus intereses inclu-yen teorıa de lenguajes formales, computacion dealto desempeno, y seguridad. Ha sido supervisorde 17 disertaciones doctorales y de mas de 100 te-sis de maestrıa. Es autor de seis libros y mas de160 trabajos cientıficos en conferencias y revistas.

Leiss ha participado en cada conferencia CLEIdesde 1992. Se involucro con LANC, para defi-nir un modelo sostenible para la conferencia. Haparticipado en numerosos comites de programa,en particular como General Program Chair de laconferencia en Quito en 2011, donde se uso porprimera vez el modelo de simposios. Ha participa-do en un ECI curso en Buenos Aires y dos ERTICcursos en Asuncion. Ademas, ha dictado cursos enSiria, Brasil, Italia, Espana, Alemania, Finlandia,Tunisia, y Marruecos. Hasta la fecha, ha dado se-minarios en 32 diferentes paises.

Leiss ha participado en la acreditacion de pro-gramas en informatica. Ha sido un ACM Distin-guished Lecturer desde 1991. Ha ocupado puestos

6

http://www.conicet.gov.ar/new_scp/detalle.php?id=22684&keywords=Santos+Rodrigo&datos_academicos=yes

http://www.diec.uns.edu.arl

cs.uh.edu

Tutorial

administrativos en su Universidad y actuado co-mo presidente de varios comites; en particular, en1994 como presidente del Senado de su universi-dad, inicio una mayor reestructuracion adminis-trativa de la Universidad de Houston. En 2014,fue premiado con la Distincion CLEI al MeritoLatinoamericano en Informatica.

Digital DisruptionJuan Pablo Guizado Serrano1

1Oracle. Lima Peruemail: [email protected]

Schedule:Fri 23th@10:15, Room: B

Existe una latente necesidad de innovacion de lasorganizaciones para garantizar su exito en la eradigital, y hacer de la disrupcion digital una opor-tunidad para la innovacion.

Tendencias como el Cloud Computing, Big Da-ta, el Internet de las cosas, redes sociales y la mo-vilidad se han revelado como los principales ejes

de la transformacion de las empresas. Por ello con-sideran que el uso de estas tecnologıas esta oca-sionando una disrupcion digital masiva en las em-presas. Debemos estar consientes que la tecnologıaaun no ha llegado a su mas alto nivel de desarrollo,al contrario estamos recien entrando al momentoen donde el despliegue correcto de esta mostrarasu verdadero valor.

Short Biography

Con casi 15 anos de experiencia en Tecnologıas dela Informacion, el Ingeniero Juan Pablo Guizadodesde hace 3 anos es Consultor en Soluciones deCapa Media en Oracle Peru, especialista en pro-ductos de Automatizacion de Proceso de negocioy Optimizacion de integracion de sistemas empre-sariales, anteriormente ha sido Administrador dePlataforma de Mision Crıtica para AplicacionesJava y Arquitecto de Aplicaciones multiplatafor-ma, conocedor ademas de construccion de siste-mas participando en multiples proyectos de en-vergadura nacional.

7

www.oracle.com

SLCGRVPI

Simposio Latinoamericano de Computacion Grafica, RealidadVirtual y Procesamiento de Imagenes (SLCGRVPI)

Indexacion de Imagenes Faciales medianteAlgoritmo basado en Permutaciones

Christian von Lucken1, Liz Gonzalez11Universidad Nacional de Asuncion. Asuncion

Paraguayemail: [email protected],

[email protected]

Schedule:Wed 21st@15:15, Room: D

In face recognition systems that work with data-bases of thousands of images it is not practical tocompare the query image against each image inthe database to determine similarity. The increa-sing use of facial recognition in various fields andlarger databases creates the need to explore me-chanisms for efficient and effective search in termsof use of computing resources and percentage ofhits. In order to ease the load of these systemsand improve response time, several alternativesto reduce or eliminate the need for the exhaus-tive search of images were developed. Indexingmethods are one of these alternatives. This paperpresents a new indexing technique based on per-mutations, which in combination with a PrincipalComponent Analysis algorithm, optimizes stora-ge of images and accelerates the search process topredict the similarity between objects. The propo-sed method shows improved behaviour when com-pared against other techniques representing thestate of the art using the FERET database.

A heuristic model for determining thesperm motility grade by video trackingDiego Garate1, Rosario Medina Rodrıguez1,

Filomen Incahuanaco1, Cesar Beltran1

1Pontificia Universidad Catolica del Peru. LimaPeru

email: [email protected],[email protected], manager [email protected],

[email protected]


Sperm motility analysis is an important part inthe fertility analysis and at the same time is agood example of the tracking multiple objects andvideo surveillance problem from a computationalpoint of view. Currently the most used methodfor this analysis is the direct inspection which isinaccurate, subjective, a not repeatable procedu-re, and difficult to teach. This article attemptsto overcome these barriers using computer vision

techniques and also proposes a heuristic model ba-sed on movement direction and Euclidean distanceto track sperm in videos obtained from an artifi-cial sperm cells simulator.

Automatic Detection of Glaucoma UsingDisc Optic Segmentation and Feature

ExtractionMaila Claro1, L. Santos1, Wallinson Silva1,Flavio Henrique1, Alcilene Sousa2, Andre

Macedo Santana11Univeridade Federal do Piauı. Brazil,

2Universidade Federal do Maranhao. Brazilemail: [email protected],

[email protected],[email protected], [email protected],

[email protected],[email protected]


Digital image processing is highlighted in the me-dical setting for automatic diagnosis of diseases.Glaucoma has no cure, and is the second leadingcause of blindness wordwide. Currently there aretreatments to prevent vision loss, but the diseasemust be discovered still in the early stages. Thus,this study aims to develop an automatic detectionmethod of Glaucoma in retinal images. The met-hodology used in the study was: image acquisition,segmentation of the optic disc area (DO) in reti-nal images, color feature extraction and entropyin the targeted area and shortly after the selec-tion of attributes. Finally, classification of imageswas conducted to determine wether or not glau-coma was present.

Revisiting the Visibility Problem with aHybrid Structure Paradigm

Icaro da Cunha1, Luiz Goncalves11Universidade Federal do Rio Grande do Norte.

Natal Brazilemail: [email protected], [email protected]

Schedule:Wed 21st@16:15, Room: B

We revisit the problem known as the visibility pro-blem, for computing a potentially visible set ofprimitives, proposing a solution that can be usedto accelerate the processing in real-time 3D visua-lization applications. We come up with a resultingdry structure in the sense of data reduction thatcan be used for on-line, interactive applications.Our main goal is to load the minimum amountof primitives from the scene during the rendering

8

www.una.py

www.pucp.edu.pe

www.ufpi.br

www.ufma.br

www.ufrn.br

SLCGRVPI

stage, as possible. For this purpose, our algorithmexecutes the culling by using a hybrid paradigmbased on viewing-frustum, back-face culling andocclusion models. Results have shown a substan-tial improvement over these traditional approa-ches if applied separately. This novel approach canbe used in devices with no dedicated processors orwith low processing power, as cell phones or em-bedded displays, or to visualize data through theinternet, as in virtual museums applications.

3D Mesh Interest Point Detection usingGISIFs and Heat Diffusion

Jan Jose Hurtado Jauregui1, Cristian Lopez delAlamo1,2, Madeley Karina Coaquira Congona1

1Universidad Nacional de San Agustın. ArequipaPeru,

2Universidad La Salle. Arequipa Peruemail: [email protected],



To facilitate processing of 3D objects it is com-mon to use high-level representations. The interestpoints are one of them. An interest point shouldpossess a distinctive feature regarding its localityand should be stable in different instances of theobject. This article proposes a descriptor basedon symmetry (GISIFs) and heat diffusion (HKS).From this features, we select a set of representativepoints. The GISIFs referenced in this article hasnot been used to extract local features. We compa-re our results with the results of other techniques,which make up the state of the art in interest pointdetection. We use a benchmark that evaluates theaccuracy of the selected points with respect to anideal set of interest points.

Deteccao Automatica do Glaucoma pelouso de Descritores Locais, Matriz GLCM

e Aprendizado de MaquinaWallinson Silva1, Maila Claro1, L. Santos1,Flavio Henrique1, Alcilene Sousa1, Nayara

Moura11Univeridade Federal do Piauı. Picos-PI Brazil

email: [email protected],[email protected],

[email protected],[email protected],


Schedule:Thu 22st@10:15, Room: D

Glaucoma is an asymptomatic disease and is con-sidered the second greater cause of blindness in

the world, it attacks the optic nerve causing irre-versible damage to the visual field. Although thereis no cure there exist treatments, which can be fa-cilitated from a early diagnosis. Thus, this workconsists of methods for feature extraction in reti-nal images from the use of descriptors for the au-tomatic detection of glaucoma. These methods areconstructed from the use of bag of words, whichare robust forms of representation, and the tex-ture descriptors Binary Pattern Place and GreyLevel Co-occurrence Matrix (GLCM).

A Texture and Curvature Bimodal LeafRecognition Model for Identification of

Costa Rican Plant SpeciesErick Mata-Montero1, Jose Carranza-Rojas11Costa Rica Institute of Technology. Cartago

Costa Ricaemail: [email protected], [email protected]


In the last decade, research in Computer Visionhas developed several algorithms to help botanistsand non-experts to classify plants based on imagesof their leaves. LeafSnap is a mobile applicationthat uses a multiscale curvature model of the leafmargin to classify leaf images into species. It hasachieved high levels of accuracy on 184 tree spe-cies from Northeast US. We extend the researchthat led to the development of LeafSnap along twolines. First, LeafSnap’s underlying algorithms areapplied to a set of 66 tree species from Costa Rica.Then, texture is used as an additional criterion tomeasure the level of improvement achieved in theautomatic identification of Costa Rica tree spe-cies. A 25.6 % improvement was achieved for aCosta Rican clean image dataset and 42.5 % fora Costa Rican noisy image dataset. In both cases,our results show this increment as statistically sig-nificant.

An Open Source Framework To ManageKinect On The Web

Francisco Moreno1, Esmitt Ramırez1, FranciscoSans1, Rhadames Carmona1

1Central University of Venezuela. CaracasVenezuela

email: [email protected],[email protected],[email protected],

[email protected]


Human-computer interaction has had continuouschanges in recent year, with a significant impro-vement in touch screens and motion sensors. New

9

www.unsa.edu.pe

www.ulasalle.edu.pe

www.ufpi.br

www.itcr.ac.cr

www.ucv.ve

SLCGRVPI

sensing technologies, like Microsoft Kinect, provi-de a low-cost way to add interactivity with gestu-res and postures. The current tendency is to deve-lop software for the web. Interactions through theKinect can be an additional benefit for these appli-cations. While there is a solution to use the Kinectin the web, it is only supported on Windows plat-form by using Internet Explorer browser. In thispaper, we propose a robust, interoperable, elegantand efficient server-client open source frameworkwhich allows interacting with the Kinect or similarcapture device from browsers. The tests supportour hypothesis, resulting in a low consumption ofmemory/time in different modern browsers. Also,an experimental test was performed to prove itsusefulness, getting a rate of 30fps successfully.

An Efficient approach for interest pointsdetection in non-rigid shapes

Cristian Lopez del Alamo1, Luciano ArnaldoRomero Calla2, Lizeth Joseline Fuentes Perez2

1Universidad de La Salle. Arequipa Peru,2Universidade Federal Fluminense. Rio de

Janeiro Brazilemail: [email protected], [email protected],

[email protected]


Due to the increasing amount of data and the re-duction of costs in 3D data acquisition devices,there has been a growing interest, in developingefficient and robust feature extraction algorithmsfor 3D shapes, invariants to isometric, topologicaland noise changes, among others. One of the keytasks for feature extraction in 3D shapes is the in-terest points detection; where interest points aresalient structures, which can be used, instead ofthe whole object. In this research, we present anew approach to detect interest points in 3D sha-pes by analyzing the triangles that compose the

mesh which represent the shape, in different wayto other algorithms more complex such as Harris3D or HKS. Our results and experiments of repea-tability, confirm that our algorithm is stable androbust, in addition, the computational complexityis O(nlogn), where n represents the number of fa-ces of the mesh.

Computerized Diagnosis of MelanocyticLesions Based on the ABCD Method

Deysi Leguizamon1, Laura Bareiro1, Jose LuisVazquez Noguera1, Diego P. Pinto-Roa1, Liza A.

Salgueiro Toledo11Universidad Nacional de Asuncion. Asuncion


[email protected], [email protected],[email protected], [email protected]


Melanoma is a type of skin cancer and is causedby the uncontrolled growth of atypical melanocy-tes. In recent decades, computer aided diagnosisis used to support medical professionals; however,there is still no globally accepted tool. In this con-text, similar to state-of-the-art we propose a sys-tem that receives a dermatoscopy image and pro-vides a diagnostic if the lesion is benign or ma-lignant. This tool is based on next modules: Pre-processing, Segmentation, Feature Extraction andClassification. Preprocessing involves the removalof hairs. Segmentation is to isolate the lesion. Fea-ture extraction is considering the ABCD dermos-copy rule. The classification is performed by theSupport Vector Machine. Experimental evidenceindicates that the proposal has 90.63 % accuracy,95 % sensitivity and 83.33 % specificity on a data-set of 104 dermatoscopy images. These results arefavorable considering the performance of diagnosisby traditional progress in the area of dermatology.

10

www.ulasalle.edu.pe

www.uff.br

www.una.py

SLIHS

Simposio Latinoamericano de Infraestructura, Hardware ySoftware (SLIHS)

Scratchpad Memory Management UsingData-Prefetching

Ivan Saraiva Silva1, Hildebrando Segundo11Federal University of Piauı. Teresina Brazil

email: [email protected], [email protected]

Schedule:Mon 19th@10:15, Room: D

Nowadays, multicore processors including dozensof cores are a reality and the challenge of de-signing microprocessors with hundreds of coresdoes not appear far from being achieved. Howe-ver, some questions remain open, despite the pro-posals found in the literature. Memory latency,data locality and programmability are some ofthe more critical ones. As the number of co-res in multi-core processor architectures increa-ses, so does the need of more efficient memoryhierarchy and resource management scheme. Thedesign of efficient memory hierarchy and memorymanagement policy becomes a very important re-search subject in multi core architecture field.The usage of resource allocation strategies beco-mes increasingly necessary. This paper presentsthe design and implementation of a multi-corenode with support to memory configuration andhardware-based scratchpad memory management.The node consists of 9 cores (8 slave cores anda master core) and a configurable local memory.The proposed architecture was implemented withVHDL and prototyped with FPGA developmentboards. Experimental results show that the pro-posed hardware-based scratchpad memory mana-gement unit can achieve 99,0 of hits in data ad-dress prediction and 97.1 % of hits when searchingdata-frames in the local memory.

Cloud Elasticity for HPC Applications:Observing Energy, Performance and CostVinicius Facco Rodrigues1, Gustavo Rostirolla1,

Rodrigo da Rosa Righi1, Cristiano Andre daCosta1, Jorge Luis Victoria Barbosa1

1Universidad de Vale do Rio dos Sinos. SaoLeopoldo Brazil

email: [email protected],[email protected], [email protected],

[email protected], [email protected]


Elasticity is one of the most known capabilitiesrelated to cloud computing, being largely deplo-yed using thresholds. In this way, limits are used

to drive resource mangement actions, leading tothe following problem statements: How can cloudusers set the threshold values to enable elasticityin their cloud applications? And what is the im-pact of the application’s load pattern on the elasti-city? This article answers these questions for ite-rative high performance computing applications,showing the impact of both thresholds and loadpatterns on application performance and resour-ce consumption. To accomplish this, we develo-ped a reactive and PaaS-based elasticity modelcalled AutoElastic and employed it over a privatecloud to execute a numerical integration applica-tion. Here, we are presenting an analysis of bestpractices and possible optimizations regarding theelasticity and HPC pair. Considering the results,we observed that the upper threshold influencesthe application time more than the lower one.

Optimal Cloud Resource Allocation byMeans of the Analytic Hierarchy ProcessValter Messias1, Julio Estrella1, Ricardo Ehlers11University of Sao Paulo . Sao Carlos-SP Brazil

email: [email protected],[email protected], [email protected]


Improving resource provisioning for applicationshosted in the cloud is a major research challenge.This is because is necessary to address two conflic-ting objectives: meet customer requirements; andsave money. As there is a delay, which can take mi-nutes, between the request for a new resource andit been ready for use, it is necessary to predict thefuture demand for each time interval. However,there is not a prediction model that is appropria-te in all cases. To resolve this problem, this paperproposes the combination of different forecastingmodels by the analytic hierarchy process. In thisway, we intend to create a generic solution, ableto optimize the allocation of resources to severalapplications types, with different demand types.The results obtained by simulation show that ourproposal achieves this goal.

A Multi-Objective Approach for VirtualNetwork Embedding

Enrique Davalos Gimenez1, Benjamın Baran1,Cristian Aceval2, Victor Franco1

1Universidad Nacional de Asuncion. San LorenzoParaguay,

11

www.ufpi.br

www.unisinos.br

www.icmc.usp.br

www.pol.una.py

SLIHS

2Universidad Nacional de Asuncion. San LorenzoParaguay

email: [email protected], [email protected],[email protected],[email protected]


Network Virtualization is a key technology for theFuture Internet, as it allows the deployment of in-dependent virtual networks using resources of thesame basic infrastructure. An important challen-ge in the dynamic provision of virtual networksresides in the optimal assignment of physical re-sources (nodes and links) to requirements of vir-tual networks. This problem is known as VirtualNetwork Embedding (VNE). For the resolution ofthis problem, previous research has focused ondesigning algorithms based on the optimizationof only one objective. On the contrary, in thiswork we present a multi-objective algorithm calledVNE-MO-ILP for solving dynamic VNE problem,which calculates an approximation of the ParetoFront considering simultaneously resource utiliza-tion and load balancing. Results of experiments,using a network simulator, probe that the propo-sed algorithm is better or at least comparable tothe state-of-the-art algorithm.

An Experimental Study on theEffectiveness of Trilateration and

Probabilistic Multilateration for PositionEstimation in MANETs

Anabel Pineda-Briseno1, RolandoMenchaca-Mendez2

1Instituto Tecnologico de Matamoros.Tamaulipas Mexico,

2Instituto Politecnico Nacional.. Mexico- D.F.Mexico

email: [email protected],[email protected]


In this paper we present an experimental studyof trilateration and probabilistic multilateration.Both methods are techniques aimed at estimatingthe location of a point, based on the position ofthree or more references and distances to them. Toasses the effectiveness of the aforementioned met-hods, we present a characterization of the variouserror conditions, a sensitivity analysis of the trila-teration and probabilistic multilateration to mea-sure the level of collinearity among references, anda simulation-based analysis describing the resultsof detailed simulation experiments used to studythe performance of DV-Hop that use trilateration

as method to estimate the position of their no-des, as well as the version of DV-Hop that useprobabilistic multilateration for the same purpo-se. DV-Hop is the most representative range-freepositioning algorithms in the context of mobile adhoc networks (MANETs). Our experimental re-sults show that the probabilistic multilaterationmethod is superior to the traditional trilaterationmethod.

gr-isdbt: An ISDB-T 1-segment ReceiverImplementation on GNU Radio

Federico Larroca1, Pablo Flores1, GabrielGomez1, Vıctor Gonzalez-Barbone1, Pablo

Belzarena11Universidad de la Republica. Montevideo

Uruguayemail: [email protected], [email protected],

[email protected], [email protected],[email protected]


Several countries in the world are undertaking animmensely challenging task: successfully performthe so-called “analogical blackout” of television. InLatin America in particular, the chosen transmis-sion scheme is, in most cases, ISDB-T (the Japa-nese standard, later adapted by Brazil). Key tothe success of this blackout is a thorough unders-tanding of the technology of choice. The presentpaper intends to be a contribution to this unders-tanding: an open, free and software-based ISDB-T1-segment receiver. Such receiver may for instan-ce be used to evaluate improvements to the stan-dard (or different algorithms for the receivers), oreven be used as a measurement tool (since one hasaccess to the whole receiving chain). In additionto presenting our implementation, we discuss thetechnology that enables the software-based recei-ver: Software Defined Radio, and in particular thesoftware framework GNU Radio.

MultiObjective Robust Network Designunder Uncertain Traffic. An approach

based on Evolutionary AlgorithmsAdolfo Arteta1, Diego P. Pinto-Roa1




Given the importance and complexity of the pro-blem of robust network design, this work, studiesthe robust network design subject to guaranteecertain level of quality of service, so that with the

12

www.cnc.una.py

www.itmatamoros.edu.mx

www.universidad.edu.uy

www.una.py

SLIHS

reservation of an adjustable bandwidth for eachnode, the network is not negatively influenced bytraffic from the rest of the network. Therefore aMultiObjective Evolutionary Algorithm (MOEA)is proposed to solve and find the robust networkdesign, inspired by the concept of price of robust-ness, which simultaneously minimizes the cost ofnetwork, inequity traffic and maximizes the trafficservice in the worst case scenario. Finally, experi-mental results show the benefits of the proposedapproach to get a set of non-dominant solutionson which it can make, a better decision.

Auction-based Resource Provisioning inCloud Computing. A Taxonomy

Sara Arevalos1, Fabio Lopez-Pires1,3, BenjamınBaran2

1National University of Asuncion. Paraguay,2East National University. Paraguay,3Itaipu Technological Park. Paraguay



Amazon Web Services marketizes its idle com-puting resources through its spot instances offer.These resources are offered through an auction-based scheme at extremely low prices. A runninginstance can be shutdown whenever the spot pricerises above the user bid. Several challenges and op-portunities emerge from this new computing para-digm. This work proposes for the first time a taxo-nomy on auction-based cloud computing resourceprovisioning, based on the study of the most re-levant literature. The studied works are classifiedaccording to: (1) provider or user perspective, (2)problem solved, (3) optimization approach, (4) ob-jective functions and (5) solution techniques. Fi-nally, new prospective research subjects are identi-fied and proposed for this promising research area.

GAIA Infrastructure: a Framework for theManagement of Information and

Communication Technology InfrastructureLuis Horacio Ramos Isique1, Rodolfo Miranda de

Barros1, Bruno Bogaz Zarpelao11Universidade Estadual de Londrina.

Londrina-PR Brazilemail: [email protected], [email protected],

[email protected]


Organizations have experienced a significant de-pendence on their Information and Communica-tion Technology (ICT) infrastructure. To mana-ge this infrastructure, they should follow seve-ral standards and good practices, which makes

this task a challenge. In this paper, we proposethe framework GAIA Infrastructure, that aims tosupport the ICT infrastructure management un-der different perspectives such as configuration,structured cabling, fault identification and miti-gation and information security. The frameworkis composed by a Diagnostic Assessment Ques-tionnaire, a Maturity Model and seven AssessmentAxes. The framework was applied in three organi-zations and, therefore, they could identify strongand weak points of their infrastructure manage-ment practices.

Protection with Quality of Service inoptical WDM networks using

Many-Objective Ant Colony OptimizationBenjamın Baran1, Julio Paciello1, Tania Nunez1,

Victor Ayala11Universidad Nacional de Asuncion. San Lorenzo

Paraguayemail: [email protected],[email protected],



This paper presents a new many-objective for-mulation of the Routing and Wavelength Assign-ment (RWA) problem in wavelength-routed opti-cal WDM networks with protection, consideringQuality of Service (QoS). A modification of theMulti-Objective Ant Colony System (MOACS) al-gorithm is proposed to solve the problem. Expe-rimental results show that the proposed approachis a promising alternative compared to the origi-nal MOACS. The proposed algorithm is also com-pared to a Mixed Integer Linear Programming(MILP) implementation proving the results ob-tained are very pretty closed, with a significantdecrease in runtime.

The Use of Different Strategies of SearchSpace Reduction in Mitigation ofOptimization Selection Problem

Nilton Queiroz Junior1, Anderson Faustino daSilva1

1Universidade Estadual de Maringa. Maringa -PR Brazil



Compiler optimizations are transformations thatare applied on the source code to improve its per-formance. Many times its a complex task to choosewhich optimizations set must be used, so, usually

13

www.una.py

www.une.edu.py

www.pti.org.py

www.uel.br

www.una.py

www.uem.br

SLIHS

there are chosen optimization levels given by thecompiler. However, these optimization levels arenot always good enough for all programs. Thus,it is needed to search for sets to improve specificprograms. Currently the Best10 algorithm is oneof the best algorithms to mitigate the optimiza-tions selection problem. This algorithm requiresone reduced search space to infer which are theoptimization sets that must be applied during thecompilation of programs. This work presents theimpact of the use of different search space crea-tion strategies used by the Best10 algorithm. Theresults shows that sophisticated strategies do notalways provide the best results.

Improved Batch Elimination: A FastAlgorithm to Identify and RemoveHarmful Compiler Optimizations

Ewerton Daniel de Lima1, Anderson Faustino daSilva1

1State University of Maringa. Parana Brazilemail: [email protected],

[email protected]


Modern compilers provide several optimizationsthat can be applied to the source code, in orderto increase its performance. Due to the complexrelationship between various optimizations, disco-vering harmful compiler optimizations is a pro-blem in the context of compilers. Strategies basedon iterative compilation try to solve this problemevaluating the performance of the compiled pro-gram using different sets. In this context, Com-bined Elimination is an efficient iterative com-pilation strategy. The purpose of Combined Eli-mination is to identify the harmful optimizationsand remove them in an iterative compilation pro-cess. Combined Elimination provides good results,which are close to those founded by an exhausti-ve search approach. However, its drawback is thenumber of program runs. In this paper, we propo-sed an iterative compilation algorithm, named Im-proved Batch Elimination. This algorithm is basedon the first step towards Combined Elimination,the Batch Elimination algorithm. The goal of Im-proved Batch Elimination is to produce results si-milar to Combined Elimination, with a complexitysimilar to Batch Elimination. In other words, thegoal is to produce good results and to be fasterthan Combined Elimination. We evaluate our al-gorithm by measuring the performance of SPECCPU 2006, POLYBENCH and CBENCH bench-marks under a set of 63 LLVM compiler optimi-zations. The results indicate that Improved BatchElimination is a good strategy to remove harmfulcompiler optimizations, using few program runs.

HTTP-WS-AD: An Anomaly Detectororiented to web applications and web

servicesJose Gimenez1, Cristian Cappo1



Schedule:Tue 20th@10:15, Room: D

Web applications have become the most popularsystems to be developed today. This is becausethey have several advantages compared to “desk-top” systems. Due to massive use of web applica-tions they have become one of the main targetsfor cyberattacks. They are also used as the princi-pal vectors for more sophisticated attacks. In thispaper we present an online anomaly based detec-tor for web applications which implements variousdetection models proposed in the literature. Ourproposed detector includes new anomaly modelsfor HTTP requests based on XML or JSON whichare formats that usually used in web services andAJAX applications. We also present a frameworkto include and to evaluate new anomaly models inthe detector.

Solving dense linear systems with hybridARM+GPU platforms

Juan Pablo Silva1, Ernesto Dufrechou1, EnriqueQuintana2, Alfredo Remon3, Peter Benner31Universidad de la Republica. Montevideo

Uruguay,2Universidad Jaime I. Castellon Spain,

3Max Planck Institute for Dynamics of complexTechnical Systems. Magdeburg Germany




The necessity of reducing the energy consumptionwhile improving the computational performancehas encouraged the development of new hardwareplatforms. In this line, hybrid architectures thatintegrate ARM processors with graphics accelera-tors offer a positive balance between computingcapabilities and energy requirements. However, inorder to make an efficient use of this hardware, itis necessary to develop new methods and compu-tational kernels, as well as to adapt existing ones.The solution of linear systems of equations is a ba-sic operation in the solution of different problems.Its relevance and computational cost has motiva-ted an important amount of work, and in con-sequence, it is possible to find high performance

14

www.uem.br

www.una.py


www.ujiapps.uji.es

SLIHS

solvers for most hardware platforms. In this workwe study the solution of dense linear systems ofequations in an NVIDIA Jetson TK1 device viathe Gauss-Huard method. The experimental eva-luation shows that the new solvers outperform theones available in the MAGMA library for systemsof dimension n <= 6000.

Network Virtualization in OpticalNetworks with Traffic Grooming

Enrique Davalos Gimenez1, Marcos Tilerıa1,Aloysius Yu1, Benjamın Baran1



[email protected]


The Virtual Optical Network Embedding pro-blem, also called VONE, deals with the efficientmapping of virtual resources onto opticals net-works. This study proposes a heuristic algorithmcalled Sd-Mapping to resolve the VONE problemusing traffic grooming techniques to carry seve-ral lower traffic requirements onto a single wa-velength. The proposed algorithm it is comparedwith an reference algorithm of the literature withthe following four metrics: Number of used wave-lengths, Weighted number of lightpaths, Averagenumber of physical hops and Grooming ports usa-ge. The experimentals results show that the pro-posed algorithm can be regarded as a valid alter-native based on promising results obtained in thefollowing metrics over three test networks: Num-ber of wavelengths, Weighted number of light-paths and Average number of physical hops. In ad-dition, this work compares five different groomingpolicies and the conclusion of this analysis is thatthe policy that minimizes the number of physical

hops (MinThp) is the preferred policy consideringthe four metrics simultaneously.

A New Approach to the MassiveProcessing of Satellite Images

Wilder Nina1, Alvaro Mamani-Aliaga2, ReneCruz1, Juber Serrano1, Jaime Cuba1, Yoni

Huaynacho1, Yessenia Yari2, Pablo Yanyachi11National University of San Agustın. Arequipa

Peru,2Catholic University of San Pablo. Arequipa

Peruemail: [email protected],

[email protected], [email protected],[email protected], gerson [email protected],



Technological advances in the field of RemoteSensing generate large volumes of geospatial da-ta. Current geographic information systems (GIS)doesn’t support the massive processing of satelli-te imagery, two examples of this kind of softwareare: (i) the Brazilian Spring project, GIS and re-mote sensing image processing system (ii) QGIS,a free and open source GIS. To achieve massiveprocessing, we have the HIPI framework, it provi-des a solution for how to store a large collection ofimages and works on the Hadoop Distributed Fi-le System. Currently, HIPI only supports specificimage formats, such as, JPEG, PNG and PPM.This article is presentes a new approach to distri-buted processing of considerable amounts of sate-llite images. We make an extension of HIPI to sup-port satellite images format, TIFF, this fact helpsto preserve the information, process and analyzesatellite images massively to have results fasterthan the traditional approach.

15

www.pol.una.py

www.unsa.edu.pe

www.ucsp.edu.pe

SLIOIA

Simposio Latinoamericano de Investigacion de Operaciones eInteligencia Artificial (SLIOIA)

Optimizacion de Enjambre de Partıculaspara Problemas de Muchos ObjetivosMateo Torres Bobadilla1, Benjamın Baran2

1Universidad Catolica “Nuestra Senora de laAsuncion”. Asuncion Paraguay,

2Universidad Nacional del Este. Ciudad del EsteParaguay



The difficulty to solve problems with many,possibly conflicting, objectives logically increaseswith the number of objectives, what makes themdifficult to solve using multi-objective algorithmslike the well known NSGA-II. Therefore, this workproposes the use of a particle swarm optimiza-tion (PSO) algorithm to solve many-objective pro-blems. The main premise of this work is thatMOPSO may be a good option for solving many-objective problems, presenting experimental evi-dence that supports this premise using the wellknown DTLZ benchmark with different perfor-mance metrics such as hypervolume, coverage andgenerational distance, among others.

HGVPRLB: a hybrid algorithm forsolving binary problems

Josiane da Costa Vieira Rezende1, Rone Ilıdioda Silva2, Marcone Freitas Souza1

1Universidade Federal de Ouro Preto. OuroBranco-MG Brazil,

2Universidade Federal de Sao Joao Del Rei.Ouro Branco-MG Brazil



In this paper it is proposed a hybrid algorithm, so-called HGVPRLB, for solving generic binary pro-blems. The algorithm HGVPRLB combines theheuristic procedures GRASP, Variable Neighbor-hood Descent, Constraint Propagation and LocalBranching Cuts. It was tested in a set of binaryproblems from MIPLIB 2010 in order to checkboth its ability to obtain feasible solutions as itsability to improve the value of these solutions var-ying the processing time. Computational experi-ments showed that when the processing time in-creases the algorithm can increase the number of

feasible solutions found in the set as well the qua-lity of the solutions. Besides it, the proposed algo-rithm outperforms another algorithm of literature,as well as two other open source solvers.

Active learning algorithms for multi-labeldata

Everton Cherman1, Grigorios Tsoumakas2,Maria Monard1

1Universidade de Sao Paulo. Sao Carlos Brazil,2Aristotle University of Thessaloniki.

Thessaloniki Greeceemail: [email protected],



The iterative supervised learning setting, in whichlearning algorithms can actively query an oraclefor labels, e.g. a human annotator that unders-tands the nature of the problem, is called activelearning. As the learner is allowed to interactivelychoose the data from which it learns, it is expec-ted that the learner would perform better withless training. The active learning approach is ap-propriate to machine learning applications whe-re training labels are costly to obtain but unla-beled data is abundant. Although active learninghas been widely considered for single-label lear-ning, this is not the case for multi-label learning,in which objects can have more than one class la-bel and a multi-label learner is trained to assignmultiple labels simultaneously to an object. The-re are different scenarios to query the annotator.This work focuses on the scenario in which theevaluation of unlabeled data is taken into accountto select the object to be labeled. In this scenario,several multi-label active learning algorithms wereidentified in the literature. These algorithms wereimplemented in a common framework and expe-rimentally evaluated in two multi-label datasetswhich have different properties. The influence ofthe properties of the datasets in the results obtai-ned by the multi-label active learning algoritm ishighlighted.

Conditional Monte Carlo withIntermediate Estimations for simulation of

Markovian systemsHector Cancela1, Leslie Murray2, Gerardo

Rubino31Universidad de la Republica. Montevideo

Uruguay,

16

www.universidadcatolica.edu.py


www.une.edu.py

www.ufop.br

www.ufsj.edu.br

www.usp.br

www.auth.gr


SLIOIA

2Universidad Nacional de Rosario. RosarioArgentina,

3Bretagne Atlantique, Campus de BeaulieuRennes. Rennes France


[email protected]


For systems that are suitable to be modelled bycontinuous Markov chains, dependability analysisis not always straightforward. When such systemsare large and complex, it is usually impossible tocompute their dependability measures exactly. Analternative solution is to estimate them by simula-tion, typically by Monte Carlo simulation. But forhighly reliable systems standard simulation cannot reach satisfactory accuracy levels (measuredby the variance of the estimator) within reaso-nable computing times. Conditional Monte Carlowith Intermediate Estimations (CMIE) is a simu-lation method proposal aimed at making accurateestimations of dependability measures on highlyreliable Markovian systems. The basis of CMIE isintroduced, the unbiasedness of the correspondingestimator is proven, and its variance is shown tobe lower than the variance of the standard esti-mator. A variant of the basic scheme, that appliesto large and highly reliable multicomponent sys-tems, is introduced. Some experimental results areshown.

A Tabu Search based heuristic for policeunits positioning

Nilson Mendes1, Andre dos Santos11Universidade Federal de Vicosa. Vicosa, MG

Brazilemail: [email protected], [email protected]


Public safety is one of most demanding areas inpublic administration, having direct consequen-ces on people welfare. Creating crime containmentstrategies or providing a fast answer to emergencysituations when they occurs is a challenge. In thispaper, we use Operations Research techniques tosolve a police units positioning problem, in orderto maximize the profit associated with police co-verage in a city. We propose a model to descri-be the problem, heuristic methods based on TabuSearch and a penalty function for infeasible so-lutions. The tests are performed using instanceswith real street network of three different cities.The results show the efficacy of the penalty met-hod, increasing the number of feasible solutions

found, the good quality of the solutions generatedby the Tabu Search, and a low convergence time,even for large instances

Improving Subjectivity Detection forSpanish Texts using Subjectivity Word

Sense Disambiguation based onKnowledge

Marco Sobrevilla Cabezudo1, Nora La SernaPalomino1, Rolando Alberto Maguina Perez11Universidad Nacional Mayor de San Marcos.

Lima Peruemail: [email protected],



In this paper, we present a Sentence-level Subjec-tivity Detection method for Spanish using Subjec-tivity Word Sense Disambiguation (SWSD) basedon Knowledge. We use a classic method of WordSense Disambiguation, using the Spanish Word-Net included in Mutilingual Central Repository3.0 and the WordNet-Pr as Knowledge base. Be-cause of the alignment between the WordNet andthe SentiWordNet, we use this latter as semanticresource annotated with polarity values to deter-mine when a word expresses subjectivity and ob-jectivity, defining subjectivity levels using a fuzzyclustering algorithm previously. Due to the fewresources focused on Sentiment Analysis for Spa-nish, the Semcor corpus was used for analyzing theattributes to be used. Finally, a Rule-based clas-sifier was created to detect subjective sentences.This method was executed over a Spanish corpus,created in this work. The results show that our ap-proach contributes positively to Subjectivity De-tection task, despite of using resources created forEnglish.

Logical Analysis of Multi-Class DataJuan Avila-Herrera1, Munevver Mine Subasi21Universidad Nacional de Costa Rica. Heredia

Costa Rica,2Florida Institute of Technology. Florida USA

email: [email protected],


Logical Analysis of Data (LAD) is a two-class lear-ning algorithm which integrates principles of com-binatorics, optimization, and the theory of Boo-lean functions. This paper proposes an algorithmbased on mixed integer linear programming to ex-tend the LAD methodology to solve multi-classclassification problems, where One-vs-All (OvA)

17

www.unr.edu.ar

www.inria.fr

www.inria.fr

www.ufv.br

www.unmsm.edu.pe

www.una.ac.cr

www.fit.edu

SLIOIA

learning models are efficiently constructed to clas-sify observations in predefined classes. The utilityof the proposed approach is demonstrated throughexperiments on multi-class benchmark datasets.

Performance metrics in multi-objectiveoptimization

Nery Riquelme1, Christian Von Lucken1,Benjamın Baran2


2Universidad Nacional del Este. Paraguayemail: [email protected],



In the last decades, a large number of metricshas been proposed to compare the performan-ce of different evolutionary approaches in multi-objective optimization. This situation leads to dif-ficulties when comparisons among the output ofdifferent algorithms are needed and appropriatemetrics must be selected to perform those com-parisons. Hence, no complete agreement on whatmetrics should be used exists. This paper pre-sents a review and analysis of 54 multi-objective-optimization metrics in the specialized literatu-re, discussing the usage, tendency and advanta-ges/disadvantages of the most cited ones in or-der to give researchers enough information whenchoosing metrics is necessary. The review processperformed in this work indicates that the hypervo-lume is the most used metric, followed by the ge-nerational distance, the epsilon indicator and theinverted generational distance.

Semantic Mining in Clusters fromSignaling Pathways Networks

Carlos Rangel1, Jose Aguilar1,2, Junior AmilcarAltamiranda Perez1

1Universidad de Los Andes. Merida Venezuela,2Universidad Tecnica Particular Loja. Loja

Ecuadoremail: [email protected], [email protected],

[email protected]


This paper describes how to semantically enrichclusters from signaling pathways networks . Thestudy is divided into two phases, the first is the de-tection of clusters in signaling pathways networks, after getting these clusters , they are passed to anextraction process of centrality within each one ,so the second phase can enrich them semantically.The centrality chosen for the case study is the

measure of closeness to other nodes, and it is whois enriched semantically in each cluster. The selec-ted case study is the signaling pathway of TGF-β,and the central nodes found were enriched withthe Gene Ontology.

A Sanction-Application Mechanismconsidering Commitment Levels in

Hierarchical OrganizationsRobert Junior1, Emmanuel Freire2, Mariela

Cortes1, Gustavo Campos11Universidade Estadual do Ceara. Fortaleza

Brazil,2Federal Institute of Ceara. Iguatu Brazil


[email protected]


The coordination and the control of autonomousagent’s behavior, in Normative Multi Agent Sys-tems, are important for the achievement of sys-tems’ global goal. However, the agents’ preferencesmay not reflect the preferences expected in the or-ganizational system specification. Then, the occu-rrence of this conflict causes a decrease in systemperformance. In this context, organizational mo-dels and their social structures (norms and sanc-tions) have been proposed to improve the agent’smanagement and agent’s behavior. However, spe-cify and apply appropriate sanctions, taking ac-count each agent’s level of responsibility, requiresan approach that considers individual and collec-tive commitments and their execution. The aimof this work is to propose a mechanism to applysanctions based on contracts and the hierarchicalorganization specification, able to be adapted indifferent application domain. The approach is ba-sed in commitment trees, contracts and sanctionsin individual and collective levels.

Dynamic Approach for a DemandResponsive Transport Service

Renan Jose dos Santos Viana1, Andre Gustavodos Santos1

1Universidade Federal de Vicosa. Vicosa - MGBrazil



Demand Responsive Transport is a flexible trans-portation service that provides transport on de-mand by a fleet of vehicles, being especially use-ful in sparsely inhabited areas, which deal witha lack of transportation service. Users formulate

18

www.una.py

www.une.edu.py

www.ula.ve

www.utpl.edu.ec

www.uece.br

www.ifce.edu.br

www.ufv.br

SLIOIA

requests specifying desired locations and times ofpickup and delivery. Given a set of requests, thevehicle routes are to be planned and scheduledin order to minimize both the number of vehiclesused and the journey durations, while respectinga set of constraints imposed by the service, thepassengers and the vehicles. In a dynamic envi-ronment, requests arrive gradually along the day,and the routes must be adjusted in real time to ac-commodate the new demand. In this work we pro-pose an approach that simulates a dynamic serviceusing mixed linear programming models to produ-ce good solutions. The obtained results for a setof instances are significative when compared to astatic approach.

Discrete Choquet integral based methodfor criteria synergy determination

Ruben Bernal1, Marcelo Karanik1, Jose IgnacioPelaez Sanchez2, Estela del Rocio Yanez

Benavides31National Technological University. Chaco

Argentina,2University of Malaga. Malaga Spain,

3University of Guayaquil. Guayaquil Ecuadoremail: [email protected],



In multi-criteria decision-making processes, a setof criteria that can act independently or have so-me kind of relation between them are considered.When this latter appears, the process cannot be asimple problem due to complexity to model suchsynergy relations. To deal with these issues, a Cho-quet integral based method has been developedto determine the most appropriate ranking of al-ternatives. Therefore, a fuzzy measure that mo-dels considerations of an expert in the problemdomain must be identified. The proposed modelis a novel and efficient algorithm for determininga non-additive fuzzy measure guided by linguis-tic attributes. Finally, an illustrative example anda comparison with other aggregation operators ispresented.

A TrueSkill approach for moviesrecommendation

Laura Cruz Quispe1, Jose Eduardo Ochoa Luna21National University of San Agustın. Arequipa

Peru,2Universidad Catolica San Pablo. Arequipa Peruemail: [email protected], [email protected]


In this work a probabilistic approach based onTrueSkill for Preference Elicitation is presented.This approach allow us to tackle the “cold start”problem because relies on a content based recom-mendation system. In addition, it is valuable forhandling high uncertainty due there is no depen-dency on the number of products and users. Theonly dependency is on ratings given by users onproducts. The proposal is highly scalable due touser preferences get richer as they are added.

Bird Species Classification UsingSpectrograms

Diego Rafael Lucio1, Yandre Maldonado eGomes da Costa1

1Universidade Estadual de Maringa. ParanaBrazil



This paper describes a system for automatic birdspecies classification based on features taken fromthe textural content of spectrogram images. Thetexture features are extracted using three of themost common texture operators described in theDigital Image Processing literature: Local BinaryPattern (LBP), Local Phase Quantization (LPQ)and Gabor Filters. Aiming to perform more fa-re comparisons, the experiments were performedover a database already used in other works pre-sented in the literature. In the classification step,SVM classifier was used and the final results weretaken using 10-fold cross validation. The experi-ments were performed over a challenger datasetcomposed of 46 classes, and the best accuracy ra-te obtained is about 77.65 %.

Forecast flows in a section of the BogotaRiver by Artificial Intelligent SystemsWilliam Moscoso1, Luis Mauricio Agudelo1

1Universidad de La Sabana, Unisabana. ChıaColombia



This article presents a comparison between twotypes of intelligent models: Artificial Neural Net-works - ANN and Adaptative Neuro-Fuzzy Inter-ference System - ANFIS, for forecasting flows in asection of Bogota (Colombia) river, looking for themost efficient. The simulation was performed inthe Matlab computer software, with data collectedby hydrological stations of the Corporacion Au-tonoma Regional of Cundinamarca (CAR), fromSeptember 2009 to October 2013. The findings

19

www.utn.edu.ar

www.uma.es

www.ug.edu.ec

www.unsa.edu.pe

www.ucsp.edu.pe

www.uem.br

www.unisabana.edu.co

SLIOIA

suggest that by using artificial intelligence modelsyou can reach a successful outcome, with Corre-lation Coefficients above 90 % (CC), Mean Abso-lute Percentage Error (MAPE) below 12 %, Con-cordance Correlation Coefficient to 84 %, six otherstatistical evaluating precision and accuracy, sug-gesting that forecasts will be labeled as good andcould think of the use of these techniques in Co-lombia.

An Application of ILS heuristic toPeriodic Vehicle Routing Problem withHeterogeneous Fleet and Fixed CostsRobert Cristian Abreu1, Jose Elias Claudio

Arroyo11Universidade Federal de Vicosa. Vicosa, MG

Brazilemail: [email protected], [email protected]


This paper addresses the Periodic Vehicle RoutingProblem (PVRP) which is a variation of the clas-sical Vehicle Routing Problem. The PVRP con-sists of assigning a combination of visiting daysto each customer, and defining the set of routesfor each day of a planning horizon. The objecti-ve of the problem is the minimization of the totallength of the routes travelled by the vehicles andthe fixed costs of using the vehicles on the time ho-rizon. To solve the problem, we present an IntegerProgramming Model and two heuristic algorithmsbased on the meta-heuristic Iterated Local Search(ILS). The mathematical model is solved by usingthe CPLEX optimization solver and optimal so-lutions are obtained for small and medium-sizedinstances of the problem. The computational ex-periments show that the heuristic algorithms de-termine high quality approximate solutions at alow computational time. The results are validatedby statistical tests.

A Variable Neighborhood SearchHeuristic for the Traveling Salesman

Problem with Hotel SelectionMarques Sousa1, Luiz Satoru Ochi1, Igor

Machado Coelho2, Luciana Brugiolo Goncalves31Universidade Federal Fluminense. Niteroi, RJ

Brazil,2Universidade do Estado do Rio de Janeiro. Rio

de Janeiro, RJ Brazil,3Universidade Federal de Juiz de Fora. Juiz de

Fora, MG Brazilemail: [email protected], [email protected],[email protected], [email protected]


This work deals with the Traveling Salesman Pro-blem with Hotel Selection (TSPHS), a variant ofthe classic Traveling Salesman Problem (TSP). Inthe TSPHS, a set of hotels can be visited in stra-tegic points of the route, dividing it in a mini-mum number of trips. Each trip must not exceeda given time limit, minimizing also the total timetraveled. The TSPHS is NP-Hard, being a genera-lization of the TSP, so the main approaches in li-terature are based in Mathematical Programmingand Metaheuristics. The metaheuristics are gene-ric heuristics capable of escaping from local op-tima, usually obtaining good quality solutions inlow computational time. It is developed a heuristicbased on Variable Neighborhood Search, compa-red with the best algorithms in literature usingclassic instances. Computational results indicatethat the proposed algorithm finds solutions withfewer trips in low computational time, with a tra-veled total time comparable to the best knownsolutions.

20

www.ufv.br

www.uff.br

www.uerj.br

www.ufjf.br

SLIS

Simposio Latinoamericano de Informatica y Sociedad (SLIS)

DICREVOA: A Proposal for the Design,Creation and Evaluation of Learning

ObjectsJorge J. Maldonado1, Jorge L. Bermeo1, Magali

Mejıa11Universidad de Cuenca. Cuenca Ecuadoremail: [email protected],


Schedule:Mon 19th@10:15, Room: C

This article describes a methodological proposalfor the design, creation and evaluation of LearningObjects (LO). This work arises from the compila-tion and analysis of several LO design methodo-logies currently used in Ibero-America. This pro-posal, which has been named DICREVOA, defi-nes five different phases: analysis, design (instruc-tional and multimedia), implementation (LO andmetadata), evaluation (from the perspective of theLO producer and consumer) and publication. Themethodology is focused not only on the teachinginexperienced but also on those having a basic un-derstanding of the technological and educationalaspects related to the design of LO, therefore, theproposal emphasizes LO design activities centeredin the Kolb cycle and the use of the ExeLearningtool in order to implement the LO core. Additio-nally, DICREVOA was used in a study case, whichdemonstrates how it provides a viable mechanismfor the LO design and implementation within dif-ferent contexts. In this paper, we present DICRE-VOA, the case study to wich it was applied, andthe obtained results.

Vulcanus: A Recommender System forAccessibility based on Trails

Ismael Cardoso1, Bruno Mota1, Jorge Barbosa1,Rodrigo Righi1

1University of Vale do Rio dos Sinos. SaoLeopoldo Brazil


[email protected]


The use of recommender systems is widespreadin several everyday systems. People are now ex-posed to different offerings, based on their inter-ests, in order to anticipate decisions. However, ina different way recommender systems for accessi-bility domain require further studies and resear-ches. In order to provide resources for people with

special needs, we developed the Vulcanus, a re-commender system designed for people with di-sabilities, offering access to resources accordinguser’s needs. The recommender system approa-ches concepts from ubiquitous computing, suchas user profiles, context awareness, trails manage-ment, and similarity analysis. Vulcanus uses twodifferent approaches: resources patterns and cate-gories patterns; where different systems can useit, for example as a Web Service. Vulcanus wasevaluated in different scenarios. These scenariosused more than 800 thousands of trails that si-mulate wheelchair users using resources along sixmonths. The obtained results show that both ap-proaches are able to provide relevant recommen-dations, providing resources that in fact can sup-port users’ needs.

Explorando o Conceito de Internet dasCoisas Sociais em um Ambiente

Universitario utilizando NFCTiago Alves1, Cristiano Andre da Costa1,Rodrigo da Rosa Righi1, Jorge Barbosa1

1Universidad de Vale do Rio dos Sinos. SaoLeopoldo Brazil

email: [email protected], [email protected],[email protected], [email protected]


Abstract The use of characteristics of smart ob-jects that have interactions features with humans,gave rise to the Internet of Things (IoT). Nume-rous derivations from this concept have been pro-posed. In this article, we focus on one of those ca-lled Social Internet of Things (SIoT). SIoT priori-tizes the relationship between smart objects, whe-re the objects can establish a connection amongthemselves without the interference of their ow-ners. The purpose of this article is to explore theconcept of SIoT in a University Campus, offeringdirect communication between intelligent devices.These devices share information based on an aca-demic criteria and preferences informed by theirowners. To evaluate the proposal, we developed acase study. The preliminary results show the via-bility of the proposal.

Web Accessibility: Study web accessibilityin public places of the Colombian StateJavier Antonio Ballesteros Ricaurte1, Jairo

Armando Riano Herrera11Universidad Pedagogica y Tecnologica de

Colombia. Tunja Colombia

21

www.ucuenca.edu.ec

www.unisinos.br

www.unisinos.br

www.uptc.edu.co

www.uptc.edu.co

SLIS



En el presente artıculo se socializan los resultadosobtenidos del proceso de analisis del nivel de ac-cesibilidad web en los sitios que fueron selecciona-dos dentro del marco de la estrategia de Gobiernoen Lınea del Estado Colombiano. Inicialmente serevisan los estandares y recomendaciones que enmateria de accesibilidad web provee la organiza-cion internacional W3C en sus dos versiones deWCAG, a continuacion se contextualiza acerca dela Estrategia de Gobierno en Lınea y de la meto-dologıa implementada para el desarrollo del pro-ceso de analisis del nivel de accesibilidad web enlos sitios seleccionados, y finalmente se presentanlos resultados obtenidos en el desarrollo de la in-vestigacion realizada.

Uso del Comercio Electronico para laVenta de Cafe Tostado de Costa Rica:

Estudio de Casos de Torrefactores de CafeCostarricenses

Francisco J. Mata1, Ariella Quesada11Universidad Nacional de Costa Rica. Heredia

Costa Ricaemail: [email protected],

[email protected]


Este artıculo presenta los resultados de un estu-dio de casos multiple exploratorio sobre el uso delcomercio electronico para la venta de cafe tostadopor parte de cuatro empresas torrefactoras cos-tarricenses. Este estudio se utiliza para contestarlas preguntas de por que utilizan estas empresasutilizan el comercio electronico y como lo hacen.Con base en los resultados obtenidos, se proponenposibles factores de exito para el uso del comer-cio electronico por parte de las empresas torrefac-toras. Aunque los resultados de este estudio nopueden ser generalizados, debido a su naturalezacualitativa, los factores propuestos pueden ser uti-lizados para guiar estudios similares en Costa Ricay en otros paıses productores de cafe. La investi-gacion contribuye al estudio del uso del comercioelectronico en paıses desarrollados, y en particularcon relacion a productos propios de estos paıses,con lo cual ayuda a entender como la tecnologıade informacion puede apoyar un mayor desarrolloeconomico y social de estos paıses.

Emancipation of Access to Wheelwright’sFinancial Information

Alejandro Sartorio1,2, Guillermo Rodrıguez1,2,Daniel Tedini2, Marcelo Vaquero2, Alejandro

Hernandez21Universidad Nacional de Rosario. Rosario

Argentina,2Universidad Abierta Interamericana. Rosario

Argentinaemail: [email protected],

[email protected], [email protected],[email protected],

[email protected]


This work directly addresses issue of access to fi-nancial information in a Community DevelopmentCommittee. The strategy involves the implemen-tation of a technical artifact direct availability tostakeholder in information according to their in-terests. Technical and procedural resources deter-mine access to information as a first class objectbelonging to a technical, methodological and con-ceptual infrastructure. Information necessary forefficient utilization through its implementation ina real case of access to community building andcollaborative information is provided.

PICTOAPRENDE: Application thatcontributes to the personal autonomy of

children and youth with Autism SpectrumDisorder in Ecuador

Johanna Tobar1, Danni De la Cruz1, PaulMejıa1, Nancy Paredes1, Andrea Cardenas1,

Edison Segovia11Universidad de las Fuerzas Armadas. Quito

Ecuadoremail: [email protected], [email protected],[email protected], [email protected],



This paper describes the PICTOAPRENDE ap-plication, which is designed to improve verbalcommunication and development of personal auto-nomy in children and youth with moderate autismspectrum disorder (ASD) in Ecuador. PICTOA-PRENDE is an application on Android platform,which provides options to children and young peo-ple to learn basic routines, emergency numbersand more. Facilitating integration into society th-rough the use of digital communicators and pic-tograms. The area of intervention presents basicconcepts related to ASD, as well as various pro-blems both in behavior and in the development ofthe beneficiary population. Finally impact measu-rement was performed in a large sample reachingimportant results for research.

22

www.una.ac.cr

www.unr.edu.ar

www.uai.edu.ar

www.espe.edu.ec

SLIS

Meneduca - Social School Network toSupport the Educational Environment

Guilherme Quirino1, Nadav Mals1, VitorGroterhorst1, Solange N. Alves de Souza1, Luiz

Sergio de Souza21University of Sao Paulo . Sao Paulo Brazil,2Centro Universitario FIEO. Osasco Brazil




Abstract— Information and CommunicationTechnologies(ICTs) have changed the way of in-teracting and how knowledge is transmitted. Inaddition, the last decade represented a revolutionin the way the society interacts, due to the inten-se use of social networks. Social networks couldalso be considered a Big Data application due tothe volume, variety and velocity of data that arecreated on them. Meneduca, a social network fo-cused on educational environments, is intended toincrease communication among teachers, studentsand parents in order to improve students’ perfor-mance. The system also aims to provide data tohelp to form groups for school tasks, based on thestudents’ personalities; for this, the Big Five Testis used. The results of the application of the BigFive allows forming working groups, whose mem-bers have heterogeneous personalities, i.e., withgreater variety of ideas and skills. In addition, tea-chers can propose different themes for academicsworks, based on the preferences of their studentscaptured from networks such as Facebook. TheMeneduca was proposed to investigate two aspectsabout social networks: (1) as a data source for ot-her applications, as recommender systems; (2) asan educational aid tool.

Time series analysis of agro-meteorologicaldata through algorithms in a scalable data

mining case: Chili river watershed,Arequipa

Melissa Abarca1, Karla Mariel FernandezFabian1, Jose Herrera Quispe1

1Universidad Nacional de San Agustin. ArequipaPeru



The paper proposes a model for predicting cli-mate change, using algorithms in mining techni-ques based on approximate data, applied to agro-meteorological data, by identifying groups search

of motifs and time series forecasting. To achie-ve the goal you work with the water balancecomponents: flow, precipitation and evaporation;also took into account the climatic variety sea-sons marked by humidity (December, January, Fe-bruary, March) and dry (other months) providingbetter to abstract sub-classification for temporarydata processing three classification techniques: li-near regression, Naive Bayes and neural networks,where the results of each algorithm are comparedwith other results. Then the mathematical met-hod of linear regression predicting water balan-ce components for a period of approximately 12months on the data of dams Pane and Fraile Wa-ter Resources in River Basin Chili, Arequipa isperformed.

Cooperative estimation of VehicularTraffic using Mobile ApplicationsJoaquın Lima1, Ruben Lopez1, Alfredo

Campuzano11Universidad Nacional de Asuncion. San Lorenzo




The increasing amount of vehicles in developingcountries, where its cities do not have a well-planned road infrastructure nor installed techno-logy to monitor traffic conditions, tends to makedifficult the daily traffic and to reduce the qualityof life. Because of this, collect and analyze traf-fic information in an inexpensive and easy mannerbecomes necessary. This paper presents the imple-mentation details of a system called Autotracks,which collects information from vehicular trafficthrough mobile devices in order to approximatethe traffic conditions in real time. A Floating CarData approach is used in combination with acti-vity recognition to monitor the location of usersin moving vehicles. The real path of the vehiclesis approximated by using a Map Matching pro-cess. Then, all the trajectories are aggregated inorder to approximate the traffic condition in thelast period of time and finally the traffic statusinformation in real time is obtained.

Quality and maturity model for open dataportals

Edgar Oviedo1, Jose Norberto Mazon1, JoseJacobo Zubcoff1

1Universidad de Alicante. Alicante Spainemail: [email protected],


Schedule:Tue 20th@10:15, Room: C

23

www5.usp.br

www.unifieo.br

www.unsa.edu.pe

www.una.py

www.ua.es

SLIS

Open Government concept is experiencing an ups-wing. Open Government is based on three con-cepts (transparency, participation and collabora-tion) that require accessing data. To provide thisaccess, Open Data Portals are being implementedaround the world by every kind of organizations,mainly in the public sector. The aim of an OpenData Portal is exposing data in such a way thatreusing is facilitated. Therefore, it is necessary todefine a quality and maturity model to evaluatethe characteristics of an Open Data Portal, con-sidering different factors that can contribute toreusing potential, such like visualization, usabi-lity, granularity, data integration, reputation, re-levancy, availability and reutilization. Also, effec-tively promoting data reusing implies setting spe-cific norms to promote standardization among ins-titutions, ministries and central governments’ of-fices into the same country. This paper presentsa formal proposal to evaluate – based in expertcriteria – the quality and the maturity of an opendata portal.

Cooperative Live Coding as aninstructional model

Antonio de Carvalho Jr11Universidade de Sao Paulo. Sao Paulo Brazil



The advances on technologies have provided manytools that inspired new instructional models. Lear-ners and instructors are experiencing a diverse en-vironment where everyone can participate fromanywhere in the world and share the same lear-ning platforms. Although we already have somemanuals, tutorials, and also MOOCs that can beuseful for people who wants to learn ComputerMusic languages, the musical interaction is not of-fered in these solutions. In this paper we presentan instructional model for computer music andlive coding based on a cooperative live coding en-vironment where participants can teach and learnthrough distributed pair programming. We alsodiscuss the fundamental ideas and the tool usedon this work during the first experiments.

Predictive model of dengue focus appliedto Geographic Information SystemsMaximiliano Baez Gonzalez1, Guillermo

Gonzalez Rodas11Universidad Nacional de Asuncion. San Lorenzo


[email protected]


Dengue is a viral disease transmitted by the ae-des aegypti mosquito. In Paraguay - South Ameri-ca, health authorities carry out entomological sur-veillance activities in order to monitor the vec-tor density in endemic and non-endemic areas th-rough techniques based on the use of traditio-nal index. Currently there are numerous methodsand most practical, efficient and economic indica-tors to determine the populations of aedes aegyptimosquito as larvitraps and ovitraps. The regiona-lized information obtained from the sampling pro-cedures can be combined with environmental, de-mographic or epidemiological information in orderto obtain detailed models that have the ability tomonitor, simulate the behavior of the vector andtherefore predict a possible outbreak of dengue.This paper presents the design and implementa-tion of a predictive model to identify outbreaks ofdengue vector infestation and the representationof its spread in a geographical information sys-tem. The model is implemented as a simulator ofthe evolutionary process of vector ecology, com-posed of a set of sub-models that seek to estimatethe rate of development, mortality, reproductionand spread of the dengue vector exposed to si-mulations of climatic variations, where the initialpopulation is generated from data obtained fromlarvitraps geographically referenced, in order togenerate enough alphanumeric and geographicalinformation to contribute to the early detectionof potential disease outbreaks.

Using the Kinect Sensor with OpenSource Tools for the Development ofEducational Games for Children on

Pre-school AgeRomina Fernandez1, Christian von Lucken1

1National University of Asuncion. AsuncionParaguay



The Kinect sensor is a device that can recogni-ze the user’s hands and body movements withoutcontrollers or marks. Although Kinect was origi-nally designed for the Xbox 360 game console,programming tools for PC appeared shortly afterits release. As a result, many papers and appli-cations explore Kinect’s technology nowadays, inareas such as robotics and medicine. Kinect canbe specially useful for pre-school education, sincethe human-computer interaction with hand mo-vements supports the concept of hand-eye coor-dination, an ability to be developed during thepre-school age. This work presents a prototypethat uses Kinect and includes educational games

24

www5.usp.br

www.una.py

www.una.py

SLIS

for pre-schoolers. This prototype was tested in aneducational center for pre-school children, and wasevaluated positively. From this experience we con-cluded that these kind of tools has a high potentialin the educational field, whereas there is a need ofinteraction standards to aid the development ofsuch kind of applications.

Sumo Sensei: Design, Implementacao eTeste com Usuarios de uma FerramentaMovel para Apoiar o Estudo de Kanjis

BasicosFabio Rocha Marques1, Bruna de Menezes1,Leonardo Cunha de Miranda1, Erica Esteves

Cunha de Miranda1, Juvane Nunes Marciano11Universidade Federal do Rio Grande do Norte.

Natal Brazilemail: [email protected],[email protected],



O estudo da lıngua japonesa por latino america-nos vem crescendo, e possui varios desafios devidoa sua forma de escrita. Kanjis sao caracteres lo-gograficos muito presentes na gramatica japone-sa e, por isso, conhece-los se torna essencial paraquem almeja aprender japones. O uso de ferra-mentas computacionais que apoiem o estudo des-se idioma e cada vez mais difundido, porem hapoucas ferramentas computacionais que ensinamjapones a partir de idiomas latinos, como espan-hol e portugues, limitando o uso para os nativosdessas lınguas. Este trabalho apresenta Sumo Sen-sei, uma ferramenta movel desenvolvida com de-sign contextualizado a uma popular luta japone-sa. A ferramenta tem o intuito de apoiar o estudode 103 kanjis basicos atraves de centenas de pala-vras de vocabulario formadas a partir dos mesmos,mesclando o aprendizado a diversao proporciona-da por elementos do jogo. A ferramenta apoia oestudo por cooperacao e competicao com partidasonline em tempo real, trazendo dinamismo com ainclusao de itens ligados a cultura do sumo, alemde possibilitar o estudo individualizado por meiodo modo treinamento contextualizado ao treinoreal do sumo. Neste artigo, ainda, apresentamosresultados de um teste realizado com o intuito deavaliar o uso da ferramenta por usuarios com di-ferentes perfis e nıveis de proficiencia na lınguajaponesa.

Karuta Kanji: Jogo Educacional paraEstudar e Praticar Vocabulario com

Kanjis da Lıngua Japonesa

Fabio Rocha Marques1, Bruna de Menezes1,Leonardo Cunha de Miranda1, Erica Esteves

Cunha de Miranda1, Juvane Nunes Marciano11Universidade Federal do Rio Grande do Norte.

Natal Brazilemail: fabio [email protected],[email protected],



Atualmente e notorio o crescimento do uso de apli-cativos em dispositivos moveis para os mais di-versos propositos, incluindo apoio ao estudo dosmais diversos assuntos. Tambem ha um crescenteuso de jogos digitais para apoiar o aprendizado deidiomas, incluindo a lıngua japonesa, ainda que amaioria destes encontre-se no idioma ingles. Es-te artigo apresenta as decisoes e estrategias peda-gogicas incorporadas ao design de Karuta Kanji,um jogo educacional desenvolvido para exercitar omaior alfabeto da lıngua japonesa – o kanji –, ten-do como lıngua nativa base o portugues, o que visabeneficiar estudantes brasileiros do idioma. O de-sign do jogo foi inspirado no popular jogo japonesde karuta. O ambiente apresenta tres modos dejogo que estimulam cooperacao, aprendizado in-dividual e competitividade. Karuta Kanji possuidiversos elementos de jogos que tem o propositode fornecer ao estudante um ambiente de estu-do dinamico mesclado ao entretenimento. Ainda,apresentamos um estudo avaliativo do jogo rea-lizado com usuarios estudantes de japones, cujosresultados proveem direcoes para o aprimoramen-to da ferramenta, do ponto de vista educacional ede entretenimento.

Katakana Star Samurai: A Mobile Tool toSupport Learning of a basic Japanese

AlphabetJuvane Nunes Marciano1, Jaime Oliveira1,Bruna de Menezes1, Leonardo Cunha de

Miranda1, Erica Esteves Cunha de Miranda11Federal University of Rio Grande do Norte.

Natal Brazilemail: [email protected],




The growing use of mobile apps in our daily acti-vities is noticed, even for work, leisure or educatio-nal purposes, in which the games have conqueredconsiderable space. Also, in the current lifestyle ofpeople in general, time is scarce, and these toolscan be very helpful, for they are always available

25

www.sistemas.ufrn.br



SLIS

since the user has a mobile device. In the contextof educational purposes, games have the advanta-ges of being fun, and rather possible to play quickmatches. This work presents Katakana Star Samu-rai, a game designed for smartphones and tablets,that supports the learning of Japanese languageproviding an environment for the student to prac-tice their knowledge, focused on a basic japanesealphabet called katakana. Also, we present a testwith users that helped us define the next steps ofthis research.

Use of a OWL ontology for creatingInteractive Learning Object

Bruno Nogueira Luz1, Rafael Santos2, AndrezaSilva Areao1, Marcos Hideyuki Yokoyama1, Luiz

Egidio Cunha2, Valeria Martins3, MarceloGuimaraes4

1Instituto Federal de Educacao, Ciencia eTecnologia de Sao Paulo. Sao Paulo Brazil,

2Faculdade Campo Limpo Paulista. Sao PauloBrazil,

3Universidade Presbiteriana Mackenzie. SaoPaulo Brazil,

4Universidade Federal de Sao Paulo. Sao Paulo

Brazilemail: nogueira [email protected],

[email protected], [email protected],[email protected], [email protected],



Interactive Learning objects (ILO) are explored inthis paper in order to provide an improvement inteaching learning process, ensuring the monitoringof students by professor through the interactionprovided by ILO. The article describes an onto-logy that defines and structure the classes of theelements of the ILO’s metadata standard. This on-tology is used as the basis for the development ofa Virtual learning environment capable of runningthese objects and at the same time offer a systemof tracking and monitoring the development of thestudent during the entire process. The ontology isvalidated through the results of the use of the ILOin the environment with a class of 20 students ofthe course of computer science.

26

www.ifsp.edu.br

www.ifsp.edu.br

www.faccamp.br

up.mackenzie.br

www.unifesp.br

SLSIGE

Simposio Latinoamericano de Sistemas de Informacion deGran Escala (SLSIGE)

CARMiCLOC: Context AwarenessMiddleware in CLOud Computing

Jose Aguilar1,2, Marxjhony Jerez1, ErnestoExposito3, Thierry Viellemur3

1Universidad de Los Andes. Merida Venezuela,2Universidad Tecnica Particular Loja. Loja

Ecuador,3Universite de Toulouse. Toulouse France

email: [email protected], [email protected],[email protected],[email protected]


Technological advances such as miniaturization ofsensors, mobile devices and increased computatio-nal capabilities have given way to Context Awa-re Applications (ACC), which constantly monitorand change depending on the context. These ap-plications require a variety of features and servi-ces, particularly on the management of context.When applications/services lack part or some ofthese services, they must be consumed from othersuppliers who can provide them via a middleware.In addition, due to the changes undergone by thecontext mean that such services consumed by theACC must be dynamic; and due to the amount ofdata involved to define the context, it may requiremore computation capacity than that provided bya mobile device. For these reason it is necessarythe use of the Cloud. In this way, to support theserequirements is proposed CARMiCLOC (ContextAwareness in Reflective Middleware Cloud Com-puting), a web-service based middleware, whichcan behave as a SaaS (Software as a Service) oras a PaaS (Platform as a Service).

Comparando ZeroMQ y RabbitMQ comotecnologıas orientadas a eventosNicolas Estrada1, Hernan Astudillo1

1Universidad Tecnica Federico Santa Marıa.Santiago Chile



Modern web apps handle huge and increasingnumbers of users and operations. A rise of event-driven architecture approach and message queuesystems provide a new alternative to face this sce-nario evidencing the lack of quantitative measure-ments comparing performance and scalability bet-ween specific message queue products. This article

proposes a prototype architecture applied in Ze-roMQ and RabbitMQ, used for measure the im-pact of (1) the number of messages over perfor-mance, and (2) the numbers of consuming nodesover scalability. The results show that for bothcriteria, the degradation threshold of ZeroMQ ishigher tan RabbitMQ, thus more scalable and fas-ter.

Case-based Reasoning for Web ServiceDiscovery and Selection

Alan De Renzis1, Martin Garriga1, AndresFlores1, Alejandra Cechich1, Alejandro Zunino2

1UNComa University. Neuquen Argentina,2Universidad Nacional del Centro de la

Provincia de Buenos Aires. Buenos AiresArgentina

email: [email protected],[email protected],[email protected],



Web Service discovery and selection deal with theretrieval of the most suitable Web Service, given arequired functionality. Addressing an effective so-lution remains difficult when only functional des-criptions of services are available. In this paper, wepropose a solution by applying Case-based Reaso-ning, in which the resemblance between a pairof cases is quantified through a similarity fun-ction. We show the feasibility of applying Case-based Reasoning for Web Service discovery andselection, by introducing a novel case representa-tion, learning heuristics and three different simi-larity functions. We also experimentally validateour proposal with a dataset of 62 real-life WebServices, achieving competitive values in terms ofwell-known Information Retrieval metrics.

Monitoring and Enforcing DataProtection Laws within an E-government

Interoperabilty PlatformAndres Echevarrıa1, Dahiana Morales1, Laura

Gonzalez11Universidad de la Republica. Montevideo

Uruguayemail: [email protected],


Schedule:Thu 22st@10:15, Room: B

27

www.ula.ve

www.utpl.edu.ec

www.univ-toulouse.fr

www.usm.cl

www.uncoma.edu.ar

www.unicen.edu.ar

www.unicen.edu.ar


SLSIGE

Public agencies are increasingly required to colla-borate with each other in order to provide high-quality e-government services. This collaborationis usually based on the service-oriented approachand supported by interoperability platforms. Suchplatforms are specialized middleware-based infras-tructures enabling the provision, discovery andinvocation of interoperable software services. Inturn, given that personal data handled by go-vernments are often very sensitive, most govern-ments have developed some sort of legislation fo-cusing on data protection. This paper proposessolutions for monitoring and enforcing data pro-tection laws within an E-government Interopera-bility Platform. In particular, the proposal ad-dresses the requirements posed by the Urugua-yan Data Protection Law and the Uruguayan E-government Platform, although it can also be ap-plied in similar scenarios. The solutions are basedon well-known integration mechanisms (e.g. En-terprise Service Bus) as well as recognized securitystandards (e.g. eXtensible Access Control MarkupLanguage) and were completely prototyped leve-raging the SwitchYard ESB product.

Chilean Virtual ObservatoryMauricio Solar1, Mauricio Araya1, Luis

Arevalo1, Ricardo Contreras3, Victor Parada2,Diego Mardones4

1Universidad Tecnica Federico Santa Marıa.Chile,

2University of Santiago de Chile. Santiago Chile,3University of Concepcion. Concepcion Chile,

4University of Chile. Santiago Chileemail: [email protected], [email protected],



This paper presents the challenges, architectureand current status of the Chilean Virtual Obser-vatory (ChiVO), which is a software infrastructu-re for accessing and processing astronomical datagenerated in Chile. As ChiVO is part of the In-ternational Virtual Observatory Alliance (IVOA),we strictly follow the protocols and standards thatthis organization produce. However, there are al-ways open challenges due to the new observatio-nal technologies and local requirements that mo-tivates research on every new virtual observatory,such as the complex data models and Big Da-ta problems that the ALMA Observatory is con-fronting. The current ChiVO prototype includesIVOA compliant services as well as new solutionsdesigned for ALMA data, all of them using mo-dern software technologies.

Business Intelligence applied to LearningAnalytics in student-centered learning

processes.Guido Riofrio1, Angel Encalada1, Daniel

Guaman1, Jose Aguilar1,21Universidad Tecnica Particular de Loja. Loja

Ecuador,2Universidad de Los Andes. Merida Venezuela


[email protected]


This work aims to evaluate the use of LearningAnalytics (LA) in Higher Education; students,and attributes such as: profile, interactions in avirtual learning environment learning - for this usebusiness intelligence paradigm in order to explo-re and exploit the data from one of the actors inthe educational process is analyzed, test scores,among others, which will contribute to their edu-cational success. In particular, this paper tries toanswer the following specific objectives: To iden-tify factors that influence the decision of a collegestudent distance learning to abandon their studiesand get the profile of potentially susceptible stu-dents from their university studies. To meet thispurpose we define two analysis tasks learning anduse a business intelligence methodology to imple-ment it.

Automating the process of buildingflexible Web Warehouses with BPM

SystemsAndrea Delgado1, Adriana Marotta1

1Universidad de la Republica. MontevideoUruguay



The process of building Data Warehouses (DW)is well known with well defined stages but at thesame time, mostly carried out manually by IT peo-ple in conjunction with business people. Web Wa-rehouses (WW) are DW whose data sources aretaken from the web. We define a flexible WW,which can be configured accordingly to differentdomains, through the selection of the web sour-ces and the definition of data processing charac-teristics. A Business Process Management (BPM)System allows modeling and executing BusinessProcesses (BPs) providing support for the auto-mation of processes. To support the process ofbuilding flexible WW we propose a two BPs level:a configuration process to support the selection

28

www.usm.cl

www.usach.cl

www.udec.cl

www.uchile.cl

www.utpl.edu.ec

www.ula.ve


SLSIGE

of web sources and the definition of schemas andmappings, and a feeding process which takes thedefined configuration and loads the data into theWW. In this paper we present a proof of conceptof both processes, with focus on the configurationprocess and the defined data.

Hacia una ontologıa de arquitectura denegocio y de procesos basada en un

modelo de negocioBernhard Hitpass1

1Universidad Tecnica Federico Santa Maria.Santiago Chile



BPM como una disciplina integrada postula quela tecnologıa tiene que seguir a los procesos y los

procesos la estrategia organizacional. La literatu-ra de BPM enfatiza la importancia y necesidad dealinear la estrategia con los procesos de negocio,pero uno de los principales problemas de quienestrabajan en una unidad de procesos es plasmar eintegrar formalmente una planificacion estrategicacon los procesos organizacionales. El estudio de laliteratura muestra que no existen guıas aceptadaso estandarizadas que orienten a los profesionales alograr este objetivo de integracion. En este traba-jo se desarrolla una propuesta de una ontologıa dearquitectura de negocio y de procesos basada enun modelo de negocio como fundamento para lagobernabilidad de BPM. La arquitectura de pro-cesos se extendera hasta el nivel de los procesosque contienen la logica de negocio que se empleaen operaciones. La ontologıa se valida por mediode un caso de estudio integrado.

29

www.usm.cl

SLISW

Simposio Latinoamericano de Ingenierıa de Software (SLISW)

Mind Maps in the RequirementsTraceability

Alessandro Silveira Duarte1, Jose AugustoFabri1, Alexandre L’Erario1, Elias Canhadas

Genvigir11Universidade Tecnologica Federal do Parana.

Parana Brazilemail: [email protected], [email protected],


Schedule:Mon 19th@10:15, Room: B

This work proposes a model of requirements tra-ceability based on mind maps. This model was de-veloped for a Software Company in southern Bra-zil. In addition to the proposed model, the mapswere applied to the process description and soft-ware features specification.

Use case technique for requirementsmodeling in distributed development

environmentsViviana Alferillo1, Maria Ines Lund1, Gerardo

Matturro21Universidad Nacional de San Juan. San Juan

Argentina,2Universidad ORT Uruguay. Montevideo

Uruguayemail: [email protected],



This paper presents the process and outcomes of asystematic literature mapping about the applica-tion of the Use Cases technique for software requi-rements specification in distributed software deve-lopment environments. The purpose is to identifywhich methods, tools and methodologies are re-ported in literature as more frequently used andhow the Use Case technique is applied when theproject team is not collocated.

A methodology to guide writing aSoftware Requirements Specification

documentHelcio A. Soares1, Raimundo Moura2

1Instituto Federal do Piauı. Piauı Brazil,2Universidade Federal do Piaui. Piauı Brazil



The Requirements Engineering (RE) is the pro-cess of defining, documenting and mantaining re-quirements and it aims to support the creationand the maintenance of the Software Require-ments Specification document (SRS). This docu-ment should be produced such way that all theparticipants can understand it. It is basis for allother activities of development and its quality isfundamental for the project success. This paperpresents a methodology for guide writing SRS do-cuments consistent, unambiguous and automatedsupport in its content. The methodology is sup-ported by ERS-EDITOR tool that uses NaturalLanguage Processing techniques to automate thesteps of the methodology. A preliminary assess-ment highlights promising results for the proposedapproach.

Collaborative Approach to Security RiskManagement Information

Maicon Balke1, Lisandra Fontoura1, Luis Alvarode Lima Silva1

1Universidade Federal de Santa Maria. SantaMaria, Brazil


[email protected]


Risk management is one of the main managementprocesses of security information since it aims toidentify, analyze, evaluate and control risks thatare due to security information. To utilize users‘experiences in this process, the utilization of colla-borative tasks allows one to exploit argumentati-ve interactions between project participants thatare involved in the development of risk manage-ment debates regarding security information. Thegoal of this paper is to propose an argumentation-based collaborative approach to deal with suchrisk management of security information. The ap-proach aims to guarantee that activities defined ina security risk management process are executedaccordingly. In it, a set of rules is proposed to en-sure that the final security risk management deba-te is complete and consistent with the argumentspresented by participants of a security softwareproject. This communication protocol is tailoredto a process of security risk management that wasparticularly defined from the ISO / IEC 27005.The protocol allows users to structure and con-trol risk discussions developed by debate partici-pants using a web-based tool called RD System.

30

www.utfpr.edu.br

www.unsj.edu.ar

www.ort.edu.uy

www5.ifpi.edu.br

www.ufpi.br

www.ufsm.br

SLISW

Through this system, a case study was developedto validate the approach proposed in this work.

Evolution of a Model-driven ProcessFramework

Wilson Padua11Federal University of Minas Gerais. Minas

Gerais Brazilemail: [email protected]


We discuss the evolution of Praxis, a model-drivenprocess framework, building on feedback fromeducational and professional applications, alongthe past fifteen years. We follow the evolutionfrom Praxis first version to the current one, dis-cussing what was introduced in each. For past andcurrent versions, we classify model improvements,discussing their nature and rationale, derived fromreceived feedback.

An MDE-Based Graphical Tool for theValidation of MySQL Replication Models

Efraın Bautista Ubillus1, Nora La SernaPalomino1

1Universidad Nacional Mayor de San Marcos.Lima Peru



At modeling level, diagramming tools such as Mi-crosoft Visio are used to design MySQL replica-tion models. However, these type of tools do notallow validating whether the MySQL replicationmodel is free of errors, showing errors if any exist.Thus, we can have erroneous documentation of theMySQL replication models. Due to the lack of thisfeature, validation is done manually, which beco-mes a tedious task, time consuming and error pro-ne. This paper proposes a MDE-based graphicalmodeling tool under the Eclipse platform for theautomatic validation of MySQL replication mo-dels. In addition, once a model has been validated,the tool is capable of generating the mysqlreplica-te commands of configuration. The results of theexperiments for the errors correction of MySQLreplication models with 25 servers demonstratethat by using the proposed tool the time is re-duced in more than 87 % compared with the toolMicrosoft Visio 2013.

Dribbling Complexity in Model DrivenDevelopment Using Naked Objects,

Domain Driven Design, and SoftwareDesign Patterns

Samuel A. Soares1, Marcius Brandao1, MarielaCortes1, Emmanuel Savio Silva Freire2

1Universidade Estadual do Ceara. FortalezaBrazil,

2Instituto Federal do Ceara. Iguatu Brazilemail: [email protected],

[email protected],[email protected], [email protected]


Systems modeling and transformations that arenecessary for code generation in the model drivendevelopment approach need to realize user interfa-ce aspects and persistence infrastructure to createexecutable software. The developer does not workjust with the domain application and also the mo-deling is more complex whenever more details areneeded in the model. Therefore, we propose a de-velopment tool where the developer just modelsthe business objects, the associations between ob-jects, and their behaviors using Domain Patternsand Design Patterns. The code is generated basedon these Design Patterns and a framework, thatimplements the architectural patterns Naked Ob-jects, has the responsibility by the infrastructure.

Comparison of Software Process Models.A Systematic Literature Review

Christian Cano1, Abraham Davila1, AndresMelgar1, Marcelo Pessoa2

1Pontificia Universidad Catolica del Peru. LimaPeru,

2University of Sao Paulo. Sao Paulo Brazilemail: [email protected],[email protected],



Nowadays, there are several software process mo-dels, which fulfill different purposes, approachesand requirements. However, this proliferation cau-ses some confusion in the industry about the be-nefits or advantages of each proposal. In this con-text, studies have been conducted to determinethe existing equivalence or the extent of coveragebetween these models having used different ap-proaches to the comparisons. This work aims topresent a study of techniques and experiences oncomparison of software process models. For thisstudy, a systematic literature review was conduc-ted in relevant databases and available documentsfinding that there are few works or experiences inthis area and it represents an aspect in softwa-re engineering that requires a higher level of re-search and development. Five different methods

31

www.ufmg.br

www.unmsm.edu.pe

www.uece.br

www.ifce.edu.br

www.pucp.edu.pe

www.usp.br

SLISW

to compare process models were found and it wasidentified that the CCT – Comparison Composi-tion Tree method is the unique method that hasa graphic representation.

Using TDD for DevelopingObject-Oriented Software - A Case StudyRamon Goncalves1, Igor Lima1, Heitor Costa11Federal University of Lavras. Lavras Brazil



Maintenance of software is accomplished tomeet the users’ needs of this software (evolu-tion/correction). But, it can become hardest if thesource code architecture is difficult to understand.Test Driven Development technique can be usedto reduce this difficulty, because it leads the deve-loper to build software with simpler source code.In this paper, this technique is employed to de-velop software whose functionality is the same aslegacy software, which was developed in a traditio-nal manner, to obtain more maintainable sourcecode. Software metrics were applied in the sour-ce code of legacy and developed software and theresults showed improvements in maintainability.

Extracting Static and Dynamic StructuralInformation from Java Concurrent

Programs for Coverage TestingRafael R. Prado1, Paulo S. L. Lopes de Souza1,George G. M. Dourado1, Simone R. S. Souza1,

Julio C. Estrella1, Sarita M. Bruschi1, JoaoLourenco2

1University of Sao Paulo. Sao Carlos-SP Brazil,2Universidade NOVA de Lisboa. Lisboa Portugal




This paper proposes novel techniques for the ex-traction of structural information from the sourcecode of Java concurrent programs for their covera-ge testing. Such techniques differ from others be-cause they consider synchronization flow amongprocesses/threads, distinct paradigms of commu-nication/synchronization, method calls and poin-ter manipulation. The structural information gat-hered from the source code is kept in a test modelbased on a Parallel Control Flow Graph (PCFG)and helps the generation of an instrumented sour-ce code, used for a future generation of trace files

and to replay the concurrent execution. The re-sults show the techniques can generate both aninstrumented code and a PCFG for Java concu-rrent programs effectively, extracting static andruntime information required for structural tes-ting.

Automated Testing Framework for MobileApplications based on User-Interaction

Features and Historical Bug InformationAbel Mendez-Porras1, Jorge Alfaro-Velasco1,

Marcelo Jenkins2, Alexandra Martınez Porras21Instituto Tecnologico de Costa Rica. Cartago

Costa Rica,2Universidad de Costa Rica. San Jose Costa

Ricaemail: [email protected], [email protected],



Mobile applications support a set of user-interaction features that are independent of theapplication logic. These features include contentpresentation or navigation features. Rotating thedevice, gestures such as scroll or zoom into screensare some examples. In this paper an automatedtesting framework for mobile applications is pro-posed. Our framework integrates user-interactionfeatures, historical bug information, and an inter-est point detector and descriptor to identify newbugs. A model of the application is automaticallycreated and explored. While the exploration of themodel is performed we introduce user-interactionfeatures and we capture images. These images arepassed to a bug analyzer to search for bugs. Thebug analyzer uses an interest point detector anddescriptor to search for areas prone to bugs inthe captured images. The use of historical buginformation is proposed to determine sequencesof events to better search bugs in applications.Preliminary results show that using the proposedtechnique is feasible to identify bugs in mobile ap-plications.

An exploratory study about cross-projectdefect prediction: impact of using differentclassification algorithms and a measure of

performance in building predictivemodels.

Ricardo F. P. Satin1, Igor Scaliante Wiese1,Reginaldo Re1

1Universidade Tecnologica Federal doParana–Campo Mourao. Parana Brazil


Schedule:Tue 20th@10:15, Room: B

32

www.ufla.br

www.icmc.usp.br

www.unl.pt

www.tec.ac.cr

www.ucr.ac.cr

www.utfpr.edu.br

www.utfpr.edu.br

SLISW

Predicting defects in software projects is a com-plex task, especially in the initial phases of soft-ware development because there are a few avai-lable data. The use of cross-project defect pre-diction is indicated in such situation because itenables reuse of data from similar projects. In or-der to find and group similar projects, this paperproposes the construction of cross-project predic-tion models using a measure of performance achie-ved through the application of classification algo-rithms. To do so, we studied the combined appli-cation of different algorithms of classification, offeature selection, and clustering data, applied to1270 projects aiming to building different cross-project prediction models. In this study we conclu-ded that Naive Bayes algorithm obtained the bestperformance, with 31.58 % of satisfactory predic-tions in 19 models created with its use. This pro-posal seems to hold promise, once the local pre-dictions considered satisfactory reached 31.58 %,against 26.31 % of global predictions.

Avaliacao da Transparencia do Sistema deCompras do Governo Brasiliero

Francisco Chaves Pinto1, Denis Silva daSilveira2, Jairo Simiao Dornelas2, Charlie Silva

Lopes21Instituto Federal Pernambuco. Jaboatao dos

Guararapes Brazil,2Universidade Federal de Pernambuco. Recife

Brazilemail: [email protected],



The process of change that is happening in thecontemporary world reaches inevitably the publicsector. The search for more effective managementto provide quality and efficient services to citi-zens has been the target to be reached by mo-derns governments. Within this perspective thestudy of information systems on public organiza-tions is strategic. In this sense, this paper aims toinvestigate transparency in the electronic systemof purchases of the Federal Government of Brazil,the ComprasNet, with the perspective of usability.In other words, this study evaluated whether thenon-functional requirement of usability has beencontemplated in the ComprasNet system interfa-ces to meet the transparency criteria. This arti-cle is designed from a qualitative exploratory ap-proach about the use of ComprasNet. To achievethe goal, semi-structured interviews were conduc-ted and a field study using the observation methodwith system users. As a result of this research, we

found that the deficiency in the ComprasNet useis caused, in most cases, by the lack of usabilityrules in their interfaces. In addition, the presentstudy suggests changes in interfaces that will im-prove the usability of the system and in its trans-parency.

Graphical and Statistical Analysis ofSoftware Evolution Using Coupling and

Cohesion Metrics - An Exploratory StudyRaul Silva1, Heitor Costa2

1University of Norte do Parana. Parana Brazil,2Federal University of Lavras. Lavras Brazilemail: [email protected], [email protected]


Developing software is expensive; thus keeping ituseful to its users is important. On the other hand,due to constant maintenance performed to meetthe changing needs of users, software undergoesdegradation of its internal structure, particularlyin coupling and cohesion. Monitoring the develop-ment of software by using some of its versions canaid a Software Engineer with relevant informationto guide your maintenance activities. In this pa-per, we present a view of the evolution of versionsof software. For this, a study was conducted in 10versions of FindBugs using coupling and cohesionmetrics calculated from VizzMaintenance and Me-tric plug-ins. In this study, we applied the Pearsonlinear correlation analysis among measurements.The result showed that there is some correlationbetween these metrics, because coupling metricsdirectly influenced the cohesion metrics, with un-desirable characteristics such as high coupling andlow cohesion compromising software quality.

Performance and Accuracy conflict inMonitoring Tools for Web Services: a case

studyJael Zela Ruiz1, Cecılia Rubira1

1University of Campinas. Sao Paulo Brazilemail: [email protected],

[email protected]


Web services have become one of the most usedtechnologies in service-oriented systems. Its popu-larity is due to its property to adapt to any con-text. As a consequence of the increasing numberof Web services on the Internet and its impor-tant role in many applications today, Web servi-ce quality is a crucial requirement and demandedby service consumers. Terms of quality levels arewritten between service providers and service con-sumers to obtain some degree of quality. The use

33

www.ifpe.edu.br

www.ufpe.br

www.unopar.br

www.ufla.br

www.unicamp.br

SLISW

of monitoring tools to control service quality le-vels is very important. Quality attributes suffervariations in their values during runtime, this isproduced by many factors such as memory leak,deadlock, race data, inconsistent data, etc. Howe-ver, sometimes monitoring tools can impact nega-tively affecting the quality of service when theyare not properly used and configured. This paperaims to show the impact of monitoring tools overservice quality, when they are not used properly.The relationship between performance and accu-racy is presented and evaluated on web services.Conflict was found between performance and ac-curacy, where performance was the most affected,because it presented a degradation in its qualitylevel during monitoring.

How to Automatically Collect OrientedObject Metrics: A Study Based on

Systematic ReviewMoshe Ribeiro1, Rodrigo Reis1, Antonio

Abelem1

1Universidade Federal do Para. Belem Brazilemail: [email protected], [email protected],

[email protected]


Aim: Getting information to automatically collectobject oriented metrics (OO metrics) in order toassist the comprehension and assessment of soft-ware products. Method: A study was developedbased on a systematic review and 37 primary stu-dies were selected from 577 papers retrieved in3 databases. Result: 177 metrics that can be au-tomatically collected were catalogued.In addition,27 of these were the most referenced. The catalo-ged metrics were classified according to the qua-lity characteristics which were related; 18 collec-tion tools have been identified. In this way, it wasconcluded that there is a set of common proce-dures for collecting OO metrics and the Java andC++ are the languages with the largest numberof tools on which is possible to extract metrics.

Sistema de Analise de Incidentes paraMelhoria Contınua

Lucas Filippi1, Sandro Jose Rigo11Universidad de Vale do Rio dos Sinos. Sao

Leopoldo Brazilemail: [email protected], [email protected]


Software quality is essential for software compa-nies maintain their competitiveness. To achievecontinuous improvement in quality, companies can

rely on various techniques and tools. The properuse of these tools can provide the company withmany benefits such as reduced costs, increasedproductivity, reduced failures, increased productreliability and customer satisfaction. This paperpresents some of these techniques or tools and sug-gests the use of them together with an applicationto help a team to reduce the time spent on supportand maintenance. The improvement proposal willbe elaborated through the PDCA cycle resultingin a system that uses quality tools with manualand automated steps. The use of the system helpsthe team to identify and prioritize improvementand preventive action opportunities through in-formation taken from the incidents reported bycustomers, fostering continuous improvement.

Factors driving the adoption of ISO / IEC29110: a case study of a small software

enterpriseAbraham Davila1, Marcelo Schneck de Paula

Pessoa21Pontificia Universidad Catolica del Peru. Lima

Peru,2Sao Paulo University. Sao Paulo Brazil



Peruvian software industry is mainly composedof micro and small enterprises and presents seve-ral problems on quality, schedule and costs. Sin-ce 2012 government supports the development ofinnovative software products using standards inorder to improve quality. Quality improvement isaffected by several factors, some contributing posi-tively and others, negatively. The objective of thisstudy is to identify which factors have influence onthe practices adoption. The research method usedis unique case study and a survey to collect com-plementary data. The studied company presentedvariations in the assessments results and duringinterviews, six most influential factors were iden-tified. The company improved their software pro-cesses with the use of some practices in a projectfunded by government and ceased some practicesafter that project. The factors that positively in-fluenced were: Experience in information techno-logy and Top management support. The negativefactors were: Competitive pressure, Perceived use-fulness, Perceived ease of use and User training.

Mejora de la monitorizacion y ejecucionde procesos de negocio con integracion y

socializacionPatricia Bazan1, Jose Martinez Garro1, Roxana

Giandini1, Javier Diaz1

34

www.portal.ufpa.br

www.unisinos.br

www.pucp.edu.pe

www5.usp.br

SLISW

1Universidad Nacional de La Plata. BuenosAires Argentina




El enfoque orientado a procesos de negocio es unaspecto relevante para las organizaciones y reci-be gran atencion de la comunidad cientıfica. Losavances tecnologicos para la incorporacion de nue-vos modelos computacionales distribuidos y de as-pectos sociales en la ejecucion de los proceso de

negocio y de las herramientas que la soportan, nohan ido a la par del crecimiento en la provisionde herramientas para automatizar la gestion deprocesos de negocios. La revision de aspectos tec-nologicos modernos en BPM – distribucion de lasactividades de los procesos e inclusion de aspectossociales relacionados con su ejecucion-, contribu-ye a mejorar los resultados en las etapas de des-pliegue, ejecucion y monitorizacion de los procesosde negocio. Este trabajo propone una especifica-cion de requisitos para un prototipo de herramien-ta de gestion de procesos que incorpore aspectosde software distribuido para enriquecer los rastrosde ejecucion, e incorporar caracterısticas socialesa su gestion.

35

www.unlp.edu.ar

SLMDI

Simposio Latinoamericano de Manejo de Datos e Informacion(SLMDI)

Semantic Recommender System for theRecovery of the Preserved Web HeritageJose Aguilar1, Claudia Leon2, Omar Portilla11Universidad de los Andes. Merida Venezuela,2Universidad Central de Venezuela. Caracas

Venezuelaemail: [email protected], [email protected],

[email protected]


This paper presents a prototype of a semantic per-sonalized recommender system for a repository ofpreserved web files. To do this, we design and im-plement a semantic repository of preserved web fi-les, containing metadata associated with each pre-served site. The knowledge stored in the metadataof the semantic repository is used for the recom-mender system, in order to give prioritized recom-mendations of the different preserved web files (orweb heritage) that meet certain search criteria.The proposed recommender also considers seman-tic associations, in order to recommend not onlythe websites matched to the search criteria, butalso semantically related.

Distributed Directory System: AHealthcare Use Case for Rural AreasAlethia Hume1, Fausto Giunchiglia1, Luca

Cernuzzi21University of Trento. Trento Italy,

2Universidad Catolica ”Nuestra Senora de laAsuncion”. Asuncion Paraguay

email: [email protected], [email protected],[email protected]


The digital content of users is commonly organi-sed in local directories representing entities fromthe real world (e.g., people, locations, organisa-tions, and events). Different representations canshow different versions, using different names torefer to the same real world entity (e.g., GeorgeLombardi, Lombardi G., Dr. Lombardi). Althoughthe data in these directories are related and caneven complement each other, there are no formallinks connecting them and allowing users to shareand search across them. In this work we propo-se a Distributed Directory System, applied to AHealthcare Use Case for Rural Areas that allowspeers: (i) to maintain full control over their data;and (ii) to find different versions of an entity based

on any name that is used in the network to refer toit. We evaluate the approach in networks of diffe-rent sizes using PlanetLab and we show promisingresults in terms of scalability.

MineraSkills: Mineracao de DadosAplicada as Vagas Anunciadas noLinkedIn Visando Definir o Perfil

ProfissionalDayane Cristine Caldeira1, Ronaldo Correia1,Rogerio Eduardo Garcia1, Danilo Eler1, Celso

Olivete11Universidade Estadual Paulista. Sao Paulo




The content posted on the online social networkshas some interesting features, such as, the widedata availability, the range of subject and cons-tant updates, however its important to know howto use these data to generate knowledge. This pa-per presents a tool that use data mining technicsto explore job posts published on LinkedIn, thegoal is to define the professional profile describedin these job posts. The keyword extraction and thegeneration of association rules was employed to dothat. The results allows a researcher to identifythe most relevant skills and the relations betweenthem.

A Proposal for Customizing Queries onXML documents based on Conditional

PreferencesAngelica Medeiros1, Valeria Soares1, Eudisley

Anjos11Universidade Federal da Paraıba. Joao Pessoa




This article presents a proposal to extent theXQuery query language called XQuery-Pref sup-porting conditional preferences. Aiming to beco-me transparent to the user executing queries writ-ten in XQuery-Pref language, we use a system ca-lled XQPref, which is responsible for elicitationof dynamic preferences and the processing of the-se custom queries. We restrict the scope of thispaper on Government Open Data that, given the

36

www.ula.ve

www.ucv.ve

www.disi.unitn.it

www.uca.edu.py

www.uca.edu.py

www.unesp.br

SLMDI

information overload, has driven the demand forsensitive techniques to solve problems associatedwith querying XML documents. The governmentopen data consist of the publication and dissemi-nation of data and public information on the Webin an open format to facilitate analysis and reuse.However, information overload has attracted theconcern of customizing the query results accordingto the needs of each user.

Towards Semantic Social NetworksRonald Chenu-Abente1, Fausto Giunchiglia1,

Luca Cernuzzi21University of Trento. Trento Italy,

2Universidad Catolica ”Nuestra Senora de laAsuncion”. Asuncion Paraguay

email: [email protected], [email protected],[email protected]


Computer-enabled social services like tagging orsharing are ubiquitous in current web applicationsthat are aimed to a group of users. These servi-ces do not only add value and new functionalitiesto their applications but also create a network ofusers and services that interconnect them to a wi-der on-line ecosystem. Currently these social net-works mainly use the vast amount of user-createdcontent, and activity logs to apply to provide re-commendations and more complex services. Thispaper presents the Social Core a social networkengine that implements semantic-based functio-nalities like semantic annotations, semantic searchsemantic-enhanced access control and user privacyprotection. The Social Core was integrated as partof the SmartCampus mobile platform, which wastested by around one hundred students, and it iscurrently being further developed as part of Eu-ropean FP7 project SmartSociety.

Prediction of Tourist Traffic to Peru byusing Sentiment Analysis in Twitter

Social NetworkRicardo Linares1, Jose Herrera1, Ana Cuadros1,

Luis Alfaro11Universidad Nacional de San Agustın. Arequipa

Peruemail: [email protected],


[email protected]


—This work involves the use of tweets, from theTwitter social network in which the users mani-fest the desire to travel to the country of Peru, tobuild a predictive tool of tourist traffic. To makethis task an automated collection of tweets usingweb crawling has been built and a Naive Bayes al-gorithm has been used for sorting tweets as part ofsentiment analysis. In the final part, we shown theresults of the application of the tool for predictingthe influx of tourists to Peru.

Speeding up the combination of multipledescriptors for different boundary

conditionsRodrigo Barroso1, Marcelo Ponciano-Silva2,

Agma Traina3, Renato Bueno11Universidade Federal de Sao Carlos. Sao Paulo

Brazil,2Instituto Federal de Educacao, Ciencia e

Tecnologia do Triangulo Mineiro. Uberada-MGBrazil,

3University of Sao Paulo. Sao Carlos-SP Brazilemail: [email protected],



Content-based complex data retrieval is becomingincreasingly common in many types of applica-tions. The content of these data is representedby intrinsic characteristics, extracted from themwhich together with a distance function allows si-milarity queries. Aimed at reducing the “semanticgap”, characterized by the disagreement betweenthe computational representation of the extractedlow-level features and how these data are interpre-ted by the human perception, the use of multipledescriptors has been the subject of several studies.This paper proposes a new method to carry outthe combination of multiple descriptors for diffe-rent boundary conditions in which the balancingis carried out in pairs, starting by the best candi-date descriptor. In the experiments, the proposedmethod achieved computational cost up to 3650times smaller than the exhaustive search for thebest linear combination of descriptors, keeping al-most the same average precision, with variationslower than 0.9 %.

37

www.disi.unitn.it



www.unsa.edu.pe

www.ufscar.br

www.iftm.edu.br

www.iftm.edu.br

www.icmc.usp.br

SLTC

Simposio Latinoamericano de Teorıa Computacional (SLTC)

On discerning strings with finite automataAbuzer Yakaryilmaz1, J. Andres Montoya2

1National Laboratory for Scientific Computing.Petropolis Brazil,

2Universidad Nacional de Colombia. BogotaColombia



We study the problem of discerning strings withdeterministic finite state automata (DFAs, forshort). We begin with a survey on the historicaland algorithmic roots of this problem. Then, wefocus on the maximum number of states that arenecessary to separate two strings of a given length.We survey the most important results concerningthis issue and we study the problem from the pointof view of some alternative models of automata.The preliminary results concerning the last issuemotivate us to formulate a conjecture stating thatDFAs can separate any pair of strings by using alogarithmic number of states. We give some evi-dence supporting our conjecture.

On the real-state processing of regularoperations and The Sakoda-Sipser

problemJ. Andres Montoya1, David Casas1

1Universidad Nacional de Colombia. BogotaColombia



In this work we study some aspects of state com-plexity related to the very famous Sakoda-Sipserproblem. We study the state-complexity of the re-gular operations, we survey the known facts and,by the way, we find some new and simpler proofsof some well known results. The analysis of thestate of art allowed us to find a new and mea-ningful notion: Real-state processing. We investi-gate this notion, looking for a model of determi-nistic finite automata holding such an interestingproperty. We establish some preliminary results,which seem to indicate that there does not existsa model of deterministic finite automata havingreal-state processing of regular expressions, but,on the other hand, we are able to exhibit a deter-ministic model of finite automata having real-stateprocessing of star free regular expressions.

A Model to Guide Dynamic AdaptationPlanning in Self-Adaptive Systems

Andres Paz1, Hugo Arboleda11Universidad Icesi. Cali Colombia



Abstract—Self-adaptive enterprise applicationshave the ability to continuously reconfigure them-selves according to changes in their execution con-texts or user requirements. The infrastructure ma-naging such systems is based on IBM’s MAPE-Kreference model: a Monitor and an Analyzer tosense and interpret context data, a Planner andan Executor to create and apply structural adap-tation plans, and a Knowledge manager to sharerelevant information. In this paper we present aformal model, built on the principles of constraintsatisfaction, to address dynamic adaptation plan-ning for self-adaptive enterprise applications. Weformalize, modify and extend the approach pre-sented in [1] for working with self-adaptation in-frastructures in order to provide automated reaso-ning on the dynamic creation of structural adapta-tion plans. We use a running example to demons-trate the applicability of such model, even in si-tuations where complex interactions arise betweencontext elements and the target self-adaptive en-terprise application.

Computing Translocation Distance by aGenetic Algorithm

Lucas da Silveira1, Jose Luis Soncco-Alvarez1,Thaynara A. de Lima1, Mauricio Ayala-Rincon1

1Universidade Federal de Goias. Goias Brazilemail: [email protected],

[email protected],[email protected], [email protected]


Translocation is a useful operation on strings withchallenging questions in combinatorics of permu-tations and interesting applications in analysis ofsequences. A translocation operation essentially isthe interchange of prefixes and suffixes among twosubstrings of a string. For the case of genomes re-presented as strings, symbols that represent genesand chromosomes are modeled as substrings of thegenomes; thus, translocation is an operation thatmodels the interaction between chromosomes insi-de a genome. The translocation distance between

38

www.lncc.br

www.unal.edu.co

unal.edu.co

www.icesi.edu.co

www.ufg.br

SLTC

two genomes is defined as the minimum number oftranslocations to convert one genome into anotherand has been proved to be a meaningful mannerof modeling the evolutive distance between orga-nisms. The particular case of unsigned genomes,those in which the orientation of the genes arenot considered, is particularly difficult, while thesigned case, in which the orientation of genes isconsidered, has been proved to be polynomiallydecidable. This paper presents an innovative Ge-netic Algorithm (GA) approach to solve the unsig-

ned translocation distance problem. A distinguis-hing feature of the proposed GA is that it usesas fitness function the translocation distance forrandomly generated signed versions of the input(that is an unsigned genome). Experiments overrandomly generated strings (synthetic genomes)showed that the proposed GA approach computesanswers that are better than those computed byan 1.5+ε-approximation algorithm, the latter alsoimplemented as part of this work.

39

SIESC

XXIII Simposio Iberoamericano de Educacion Superior enComputacion (SIESC)

Philosophy of Computer Science and ItsImpact on Education: Towards the

construction of an interdisciplinary teamSylvia da Rosa1, Federico Gomez2, Alejandro

Chmiel31Universidad de la Republica. Montevideo

Uruguay,2Universidad de la Empresa. Montevideo

Uruguay,3Universidad de la Republica. Montevideo

Uruguayemail: [email protected], [email protected],

[email protected]

Schedule:Wed 21st@10:15, Room: C

This article presents an interdisciplinary experien-ce that brings together two areas of computerscience; didactics and philosophy. As such, the ar-ticle introduces a relatively unexplored area of re-search, not only in Uruguay but in the whole LatinAmerican region. The reflection on the ontologicalstatus of computing, its epistemic and educationalproblems, as well as their relationship with tech-nology, allows us to elaborate a critical analysisof the discipline and a social perception of it as abasic science.

Adoption alternatives of academicinnovations in computer science schools in

PeruJorge Alvarado1

1Universidad Nacional Mayor de San Marcos.Lima Peru



Peru has a variety of names to describe universitydegrees in computer science or related fields. Theadoption of international standards both in thearea of knowledge and professional profiles gene-rates all kinds of debates and reactions with fewadoption results. The present study introduces theinnovation diffusion theory to the analysis of thecharacteristics of the Peruvian university educa-tional system and proposes the use of these tech-niques to identify the elements necessary to allowbetter adoption by institutions. Adoption alterna-tives that fit the different types of university arepresented.

Educational Web Tool for Digital ImageProcessing

Martin Poletti1, Horacio Legal1, Jacques Facon2,Claudio Barua1, Jose Luis Vazquez1


2Pontifıcia Universidade Catolica do Parana.Brazil

email: [email protected], [email protected],[email protected], [email protected],

[email protected]


Due to its versatility, the image processing areaoffers a very wide range of techniques to solve cha-llenges in an effective way linked to fields such asmedicine, agriculture, biology, industrial automa-tion and document processing. Therefore a correctand advanced training of professionals in this areais an important task. In this sense, a new edu-cational image processing tool is currently beingdeveloped at the Facultad Politecnica of the Uni-versidad Nacional of Asuncion. The developmentfocused on improving the interaction between stu-dents and teachers and also showing new advancesin the digital image processing area. To achievethis goal, a tool is being developed to be expan-ded in the future to be adapted to new challengesand different audiences. The first stage of develop-ment was completed, which allowed developing anextendable basic tool in the near future.

Student Understanding of the C++Notional Machine Through Traditional

Teaching with Conceptual Contrapositionand Program Memory Tracing

Jeisson Hidalgo-Cespedes1, Gabriela Marın1,Vladimir Lara1

1Universidad de Costa Rica. Costa Ricaemail: [email protected],



In order to learn a programming language, a co-rrect understanding of its notional machine ismandatory. Students acquire that comprehensionmainly through visual and verbal explanationsprovided by professors, books, videos, and otherinstructional materials. This research applied theconceptual contraposition technique and programmemory tracing technique to the prevalent tea-ching method in our country: the lecture. Theunderstanding of the C++ notional machine was

40


www.ude.edu.uy


www.unmsm.edu.pe

www.una.py

www.pucpr.br

www.ucr.ac.cr

SIESC

evaluated on students of a Programming II (CS2)course that implemented the mentioned methods.Results revealed that these techniques applied tothe lecture are insufficient to help students deve-lop satisfactory mental models of the C++ notio-nal machine.

Academic Performance of UniversityStudents and its Relation with

EmploymentLaura Lanzarini1, Marıa Emilia Charnelli1,

Javier Dıaz11Universidad Nacional de La Plata. La Plata,

Buenos Aires Argentinaemail: [email protected],



Educational Data Mining collects the various met-hods that allow extracting novelty and useful in-formation from large data volumes in educationalcontexts. This paper describes the process usedto, through Data Mining techniques, identify themost relevant characteristics in relation to studentacademic performance at the School of ComputerScience of the National University of La Plata.The results obtained using the proposed methodto process the information relating to regular andnon-regular students at the UNLP allowed esta-blishing interesting relationships in relation to stu-dent academic performance. Based on the obtai-ned models it can be said that the fact that the

student works does not mean that their academicperformance decreases and young students thattake several years to join the faculty have betterperformance if they express interest in getting ajob.

University Bonding with ProductiveSector Companies: A literature review

Sandra Cabrera Alzate11Universidad Nacional de Costa Rica. Heredia

Costa Ricaemail: [email protected]


The bonding of productive sector companies anduniversities is a need that has been identified fordecades. The interplay between academia and theproductive sector, is becoming in an aspect in-creasingly important in our new demanding andchanging society. Knowledge as a strategy to crea-te competitive companies and universities as anagency that generates knowledge and preparesprofessionals to meet current demands. This do-cument focuses on an extensive literature reviewof several American authors, Europeans, Asians,Latin Americans and others, in order to identifysuccessful processes and interaction models thatcan be adopted and adapted to the area of vo-cational training in Technology information andCommunication in the context of Latin Ameri-ca.This study will serve as a starting point forfurther diagnosis and as an input for generating alinkage program in the area of computer careers.

41

www.unlp.edu.ar

www.una.ac.cr

WNAPC

V Workshop en Nomenclatura y Acreditacion en Programasde Computacion (WNAPC)

Caso Colombia: La investigacion comofactor determinante en el Aseguramiento

de la Calidad de las Instituciones deEducacion SuperiorRichard Aroca Acosta1

1Universidad Autonoma del Caribe. Colombiaemail: [email protected]


Caso Colombia: La investigacion como factor de-terminante en el Aseguramiento de la Calidad delas Instituciones de Educacion Superior

Presentaciones de la problematica dediversos paıses de la region (parte2)Ernesto Cuadros-Vargas1, Ariel Sabiguero

Yawelak2

1Universidad Catolica San Pablo. Peru,2Universidad de la Republica. Montevideo

Uruguayemail: [email protected], [email protected]


Presentaciones de la problematica de diversos paı-ses de la region (parte2)


Yawelak2






Yawelak2





Discusion y resultadosErnesto Cuadros-Vargas1, Ariel Sabiguero

Yawelak2




Discusion y resultados de la problematica existen-te en la region. Propuestas de nuevas estrategiaspara mejorar y llevar la region a un nivel mas com-petitivo en computacion tomando como referencialos ultimos avances en el area a nivel mundial.

42

www.uac.edu.co

http://socios.spc.org.pe/ecuadros

cs.ucsp.edu.pe



cs.ucsp.edu.pe



cs.ucsp.edu.pe



cs.ucsp.edu.pe


LAWCC

VII Congreso de la Mujer Latinoamericana en laComputacion (LAWCC)

Participacao Feminina em Pesquisa naPlataforma Lattes no Brasil

Maria Carolina Monard1, Ana Maria Monteiro21University of Sao Paulo. Sao Carlos-SP Brazil,2Faculdade Campo Limpo Paulista. Sao Paulo


[email protected]


The Lattes Platform, named after the Brazilianphysicist Cesar Lattes, is the major scientific in-formation system to manage information on Scien-ce, Technology, and Innovation (ST&I) related toindividual researchers and institutions working inBrazil. The Lattes platform allows managing cu-rricular information of researchers and institutionsworking in Brazil based on the so called LattesCurriculum. As the registration in the platform ismandatory to require funding from funding agen-cies, its use is disseminated throughout the Natio-nal ST&I system.This work addresses the womenin science issue in Brazilian institutions, as wellas their leadership role in Research Groups.To thisend, data related to researchers holding a PhD wascollected from the Lattes Platform and was analy-zed. It was observed that women participation inscience is similar to men participation (47 %). Ho-wever, there are differences whenever the main re-search area is taken into account. Nevertheless,women participation in Research Groups, as wellas their participation as leaders, has increased overthe years.

A Plataforma Lattes, que leva o nome do fısicobrasileiro Cesar Lattes, e o principal sistema paragerenciar informacao sobre Ciencia, Tecnologia eInovacao (CTI) no Brasil.As fontes de informacaofundamentais sao a base de Currıculos Lattes (cu-rrıculos nos padroes da plataforma) e o Diretoriode Grupos de Pesquisa. O cadastro de pesquisa-dores na plataforma e obrigatorio para os pedidosde fundos nas agencias de fomento e seu uso es-ta disseminado em todo o sistema brasileiro deCTI. Neste trabalho foram analisados alguns as-pectos da participacao da mulher no contexto dapesquisa realizada no Brasil utilizando dados co-letados da Plataforma Lattes. Foi observado quea participacao da mulher e quase a mesma que ado homem, embora existam diferencas no que serefere a distribuicao nas areas de pesquisa. Contu-do, a participacao da mulher tanto em grupos de

pesquisa quanto na lideranca desses grupos tem-seincrementado ao longo dos ultimos anos.

Digital Equity and Gender Issues in LatinAmerica

Gabriela Marın1

1Universidad de Costa Rica. Costa Ricaemail: [email protected]


Latin America and the Caribbean is a middle-income region, with the majority of its 42 coun-tries and territories belonging to that category.However, it is a heterogeneous region, rangingfrom low income countries, as Haiti, to countrieswhich have higher income and are regarded as mo-re developed, like Chile, Mexico, Argentina andBrazil. According to SEDLAC, 34.3 % of the LatinAmerican population is in the middle class (withincomes between $10 a day and $50 a day), and25.3 % are still under the moderate poverty line of$ 4 a day.

In the last decades, the region has workedto achieve the Millennium Development Goals(MDGs): poverty has been reduced to lower le-vels, more girls are in school, child mortality hasdropped, and diseases are being fought. However,gender issues remain. Too many women still diein childbirth, and more needs to be done to boostgender parity in employment and decision ma-king, as well as, access to education and repro-ductive health services. Inequality remains a keyproblem. Progress in the Region has been weakeramongst women, youth, indigenous peoples, afro-descendants and rural populations.

Digital divide is rooted in the very issues thatconstrain Latin America’s overall economic deve-lopment - income inequality, lack of infrastructureand a still-nascent technological knowledge baseaccording to UNDP. The region, as a whole, suf-fers from a poor legal framework for the develop-ment of the ICT sector, heavy administrative bur-dens, almost in-existing government prioritizationfor ICT development, low Internet penetration ra-tes, and pervasive brain drain which underminesthe potential for faster growth of the economies’ICT sectors.

An overview on gender inequality on accessto technology tools, computers, Internet and edu-cation in the region is presented. Moreover, wo-men’s low participation on the ICT sector is depic-ted. Cultural, professional and technological ba-

43

www.icmc.usp.br

www.faccamp.br

www.ucr.ac.cr

LAWCC

rriers imposed on women participation are analy-zed, and some possible actions to reduce such gen-der biases are proposed. Latin American initiati-ves to try to promote gender digital equity andthe presence of women on the ICT sector are alsodescribed.

Acciones a futuro en Latinoamerica -discusion general

Andrea Delgado1, Raquel Patino21Universidad de la Republica. Montevideo

Uruguay,2Universidad Catolica San Pablo. Arequipa Peruemail: [email protected], [email protected]


De acuerdo al avance de tecnologıa, diversos pro-fesionales van desarrollando cualidades y talentosque ayuden a contribuir con el crecimiento actual.Hoy en dıa, es importante discutir las cualidadestanto de hombres como de mujeres de apostar porintroducirse en este mundo tecnologico. Este es-pacio estara dedicado a una discusion abierta en-tre todas y todos quienes asistan a la sesion, paraintercambiar ideas y coordinar iniciativas respec-to a acciones que puedan impulsarse a futuro enLatinoamerica para promover una mayor equidadde genero y una mayor participacion de las muje-res en la informatica latinoamericana, tanto en laacademia como en la industria.

44


www.ucsp.edu.pe

CLTM

XXII Concurso Latinoamericano de Tesis de Maestrıa(CLTM)

Espelho virtual interativo para simulacaode maquiagem

Filipe Morgado Simoes de Campos1, Carlos H.Morimoto1

1Universidade de Sao Paulo. Sao Paulo-SPBrazil



Um simulador de maquiagem pode ser utilizadopara criar um ambiente virtual interativo capazde tornar o processo de escolha e experimentacaode maquiagens algo rapido, barato e flexıvel. Pa-ra seu sucesso, tal simulador deve respeitar quatrocriterios, a flexibilidade na aplicacao dos produtos,processamento em tempo real, realismo da simu-lacao e baixo custo. O projeto desenvolvido podecontribuir com uma proposta de um modelo deinteracao para um simulador de maquiagem e umalgoritmo para a simulacao de base, batom e som-bra, ambos respeitando as quatro caracterısticascitadas anteriormente.

Image Segmentation by Image ForestingTransform with Non-smooth Connectivity

FunctionsLucy A. Choque Mansilla1

1University of Sao Paulo. Sao Paulo-SP Brazilemail: [email protected]


Image segmentation, such as to extract an objectfrom a background, is very useful for medical andbiological image analysis. In this work, we proposenew methods for interactive segmentation of mul-tidimensional images, based on the Image Fores-ting Transform (IFT), by exploiting for the firsttime non-smooth connectivity functions (NSCF)with a strong theoretical background. The new al-gorithms provide global optimum solutions accor-ding to an energy function of graph cut, subject tohigh-level boundary constraints (polarity and sha-pe), or consist in a sequence of paths’ optimiza-tion in residual graphs. Our experimental resultsindicate substantial improvements in accuracy in

relation to other state-of-the-art methods, by allo-wing the customization of the segmentation to agiven target object.

Solving the Art Gallery Problem: APractical and Robust Method for Optimal

Point Guard PositioningDavi Colli Tozoni1

1University of Campinas. Sao Paulo Brazilemail: [email protected]


This Master’s thesis focused on studying and de-veloping techniques for optimally solving the ArtGallery Problem (AGP), one of the most investi-gated problems in Computational Geometry. TheAGP, which is a known NP-hard problem, con-sists in finding the minimum number of guardssufficient to ensure the visibility coverage of anart gallery represented as a polygon. We studiedhow to apply Computational Geometry conceptsand algorithms as well as Integer Programmingtechniques in order to solve the AGP to optima-lity. This work culminated in the creation of anew algorithm for the AGP, whose idea is to ite-ratively generate upper and lower bounds for theproblem through the resolution of discretized ver-sions of the AGP. The algorithm was implementedand tested on more than 2800 instances of diffe-rent sizes and classes of polygons. The techniquewas able to solve in minutes more than 90 % ofall instances considered, including polygons withthousands of vertices, greatly increasing the set ofinstances for which exact solutions are known. Tothe best of our knowledge, in spite of the exten-sive study of the AGP in the last four decades,no other algorithm has shown the ability to solvethe AGP as effectively as the one described here.For illustration, in a direct comparison with thealgorithm by Kroller et al., considered one of themost promising techniques for the AGP, our met-hod solved almost 32 % more instances than itscompetitor.In addition, we provide a free versionof our code and of our benchmark for download,which is unprecedented in the literature.

45

www.ime.usp.br

www.icmc.usp.br

www.unicamp.br

CLTD

I Concurso Latinoamericano de Tesis de Doctorado (CLTD)

Multivariate Investigation of NP-HardProblems: Boundaries BetweenParameterized Tractability and

IntractabilityUeverton dos Santos Souza1

1Universidade Federal Fluminense. Rio deJaneiro-RJ Brazil



The main goal when using computing to solve aproblem is to develop a mechanism to solve it ef-ficiently. In general, this efficiency is associatedwith solvability in polynomial time. The theoryof NP-completeness was developed to show whichproblems probably do not have polynomial ti-me algorithms. However, many NP-hard and NP-complete problems must still be solved in practice;therefore it is natural to ask if each of these pro-blems admits an algorithm whose non-polynomialtime complexity is purely a function of some sub-set of its aspects. Questions about the existence ofsuch algorithms are addressed within the theoryof parameterized computational complexity deve-loped by Downey and Fellows.

In this thesis we present a multivariate inves-tigation of the complexity of some NP-hard pro-blems, i.e., we first develop a systematic comple-xity analysis of these problems, defining its sub-problems and mapping which one belongs to eachside of an “imaginary boundary” between poly-nomial time solvability and intractability. Afterthat, we analyze which sets of aspects of theseproblems are sources of their intractability, thatis, subsets of aspects for which there exists an al-gorithm to solve the associated problem, whosenon-polynomial time complexity is purely a fun-ction of those sets. Thus, we use classical and pa-rameterized computational complexity in an alter-nate and complementary approach, to show whichsubproblems of the given problems are NP-hardand latter to diagnose for which sets of parame-ters the problems are fixed-parameter tractable,or in FPT.

This thesis exhibits a classical and parameteri-zed complexity analysis of different groups of NP-hard problems. The addressed problems are di-vided into four groups of distinct nature, in thecontext of data structures, combinatorial games,and graph theory: (I) and/or graph solution andits variants; (II) flooding-filling games; (III) pro-blems on P3-convexity; (IV) problems on inducedmatchings.

Dynamic Composition of REST servicesJesus Bellido1

1Pontificia Universidad Catolica de Chile.Santiago de Chile Chile



Service composition is one of the principles ofservice-oriented architecture; it enables reuse andallows developers to combine existing services tocreate new services. Dynamic composition requi-res that service components are chosen from aset of services with equal or similar functiona-lity at runtime. The adoption of the REST ser-vices in the industry has led to a growing num-ber of services of this type, many with similarfunctionality. The existing dynamic compositiontechniques are method-oriented whereas REST isresource-oriented, and considers only traditionalservices. The REST architectural style has attrac-ted a lot of interest from the industry due to thenon-functional properties it contributes to Web-based solutions. In this thesis, we contribute to thearea of web service composition in REST by pro-posing three techniques oriented to improve staticand dynamic composition of this type of service.First we introduce a technique for static compo-sition proposing a set of fundamental control flowpatterns in the context of decentralized composi-tions of REST services. In contrast to current ap-proaches, our proposal is implemented using theHTTP protocol and takes into account REST ar-chitectural principles. Afterwards, we present atechnique to improve the dynamic composition insecurity domain extending ReLL to ReLL-S andallowing a machine-client to interact with securedresources, where security conditions may changedynamically. Finally, we propose SAW-Q, an ex-tension of Simple Additive Weighting (SAW), asa novel dynamic composition technique that fo-llows the principles of the REST style. SAW-Qmodels quality attributes, in terms of response ti-me, availability and throughput, as a function ofthe actual service demand instead of the traditio-nal constant values. Our results validate our mainhypotheses indicating improvements with respectto alternative state-of-the-art methods. This alsoshows that the ideas presented in this thesis repre-sent a relevant contribution to the state-of-the-artof REST service compositions.

Graph Laplacian for Spectral Clusteringand Seeded Image Segmentation

Wallace Casaca1

46

www.uff.br

www.uc.cl

CLTD

1University of Sao Paulo. Sao Carlos-SP Brazilemail: [email protected]


Image segmentation is an indispensable tool to en-hance the ability of computer systems to performelementary cognitive tasks such as detection, re-cognition and tracking. In particular, interactivealgorithms have gained much attention lately, spe-cially due to their good performance in segmen-ting complex images and easy utilization. Howe-ver, most interactive segmentation methods relyon sophisticated mathematical tools whose effec-tiveness strongly depends on the kind of image tobe processed. In fact, sharp adherence to the con-tours of image segments, uniqueness of solution,high computational burden, and extensive userintervention are some of the weaknesses of most

existing techniques. In this work we proposed twonovel interactive image segmentation techniquesthat sort out the issues discussed above. The pro-posed methods rely on Laplace operators, spec-tral graph theory, and optimization approaches to-wards enabling highly accurate segmentation toolswhich demand a reduced amount of user interac-tion while still being mathematically simple andcomputationally efficient. The good performanceof our segmentation algorithms is attested by acomprehensive set of comparisons against repre-sentative stateof-the-art methods. Indeed, quali-tative and quantitative results obtained from well-known image benchmarks show that our metho-dologies outperform others. As additional contri-bution, we have also proposed two new algorithmsfor image inpainting and photo colorization, bothof which rely on the accuracy of our segmentationapparatus.

47

www.icmc.usp.br

Graduate

Collaboration Networks in Computing (Graduate)

USP Brasil: Redes de Colaboracao naArea de Computacao: perspectivas e acoes

Adenilso Simao11University of Sao Paulo. Sao Carlos-SP Brazil



Redes de Colaboracao constituem um mecanis-mo efetivo e essencial para o desenvolvimento desolucoes efetivos e de impacto para a socieda-de. Nesta palestra promovida pelo ICMC-USP,por meio das Comissoes de Relacoes Internacio-nais, de Pesquisa e de Pos-graduacao, pretende-semotivar redes de colaboracao em pesquisa, even-tualmente explorando-se tambem a colaboracaoacademia-industria, na America Latina. Entende-se que as articulacoes poderiam ocorrer no am-bito do CLEI - centro e evento, que ha decadasdesenvolve atividades relevantes para a comunida-de latino-americana. Essa colaboracao deve oco-rrer com o envolvimento de todos os agentes do sis-tema de ensino e pesquisa dos diversos paıses. Co-mo ponto central sera apresentada uma chamadapara realizacao de pos-doutoramento no ICMC-USP, visando a promover e fortalecer a coope-

racao sul-sul. Outras iniciativas e propostas seraoapresentadas e discutidas: programa de doutora-do sanduiche; programa de professor visitante bi-lateral; dupla titulacao entre instituicoes. Tem-secomo objetivo de fundo promover a mobilidadede pesquisadores na America Latina. Um facili-tador seria viabilizar um exame para ingresso empos-graduacao na america latina coordenado pe-lo centro CLEI, o que facilitaria a mobilidade dosnovos talentos para programas de mestrado e dedoutorado.

UChile Chile: Programa de Doctorado enCiencias de la Computacion en la

Universidad de ChileGonzalo Navarro1



Describire el programa de doctorado que ofrece-mos en el Departamento de Ciencias de la Compu-tacion de la Universidad de Chile, y los mecanis-mos de becas que existen para quienes postulandesde el extranjero.

48

www.icmc.usp.br

http://www.dcc.uchile.cl/~gnavarro

http://www.dcc.uchile.cl

Google

Premios de Investigacion de Google para America Latina(Google)

Learning Dynamic Action Units forThree-dimensional Facial Expression

RecognitionPablo Arbelaez1

1Universidad de Los Andes. Bogota Colombiaemail: [email protected]

Schedule:Thu 22st@10:15, Room: C

Automated understanding of facial expressions isa fundamental step towards high-level human-computer interaction. The goal of this proposalis to develop a solution to this problem by takingadvantage of the color, depth and temporal in-formation provided by an RGB-D video feed. Weplan to model human facial expressions throughthe analysis of temporal variations in the patternof activations of their natural constituents, three-dimensional Action Units. As starting point for al-gorithm development, we propose to build on ourprior experience developing convolutional neuralnetwork architectures for fine-grained localization,RGB-D scene understanding and video analysis.

Gradual Security Typing for the WebRaimil Cruz Concepcion1



How to develop secure applications for the Web?Can we help programmers write applications thatdo not compromise privacy by leaking confidentialinformation on public channels, either accidentallyor due to malicious third-party code? Researchin programming languages for security has explo-red two complementary approaches for controllingflow of sensitive data in programs: either stati-cally, with a type system, or dynamically, witha runtime monitor. A static approach is prefera-ble but might impose too much of a burden onprogrammers. Just like static and dynamic typingcan be reconciled with a gradual type system, wepropose to develop a gradual security type sys-tem, which allows programs to be partially typed,ensuring sound interoperability between the secu-rity typed and untyped fragments. Programmerscan then start with a standard program, develo-ped without security typing information, and pro-gressively evolve their system towards a programwith mixed static and dynamic guarantees. Af-ter addressing the language design issues involved

in creating a practical secure gradually-typed lan-guage, we will extend Dart, a recent programminglanguage for the Web developed by Google. Dartis a great target due to its increasing adoptionfor Web programming, and for its unique, highlypragmatic approach to type systems. The designand implementation of gradual security typing inDart will be validated by securing existing Dartapplications, and by studying how programmersuse such a gradual type system.

Interconnected Dual Biosensor for Type IIDiabetes Mellitus

Mathieu Hautefeuille11Universidad Nacional Autonoma de Mexico.

Mexico City Mexicoemail: mathieu [email protected]


The main objective of the present project is to de-velop a biosensor for diagnostic and monitoring oftype 2 diabetes mellitus. This platform will inte-grate transducers and electronics to measure glu-cose and insulin levels in real time, providing datauniversally accessible and useful for further analy-sis by physicians and clinicians, as well as for sta-tisticians for a broader screening of this disease inMexico and in certain local populations.

Urban Coordination of AutonomousVehicles

Jorge Zapotecatl11Universidad Nacional Autonoma de Mexico.

Mexico City Mexicoemail: [email protected]


The massive deployment of autonomous vehiclesin the coming years offers a great potential forimproving urban mobility. An open question ishow to coordinate selfdriving cars at intersections,which we will address in this project. Since trafficconditions change constantly, a promising alterna-tive to develop adaptive coordination mechanismsis to use self-organization. We have already de-veloped self-organizing algorithms for traffic lightcoordination with significant performance impro-vements over traditional coordination methods.We will follow a similar path exploring how auto-nomous vehicles can self-organize at intersectionsto maximize efficiency and safety without the needof traffic lights.

49

https://ingbiomedica.uniandes.edu.co

https://www.dcc.uchile.cl

http://www.fciencias.unam.mx

http://www.iimas.unam.mx

Google

50

Documents

CLEI2015 XLI Conferencia Latinoamericana de Informa tica ... · i Abstracts of the XLI Conferencia Latinoamericana de Informatica Arequipa, Peru , Octubre/Outubro/October 19th-23rd