Summary: This technical report contains the abstracts of 13 papers whose articles were authorized to be published in the XIV Workshop of Theses, Dissertations and Scientific Initiation Papers (WTD), from the Institute of Computing (IC) of the State University of Campinas (Unicamp), edition 2019 The XIV Workshop took place between 27 and 28 November 2019 and had about 137 participants, including listeners, presenters of work and organizers. There were 42 oral presentations, 51 posters, 32 lightning talk and 8 demonstrations. Students were given the possibility to choose the form of presentation (oral, poster or lightning talk), as well as to choose whether or not they wanted to publish their work in the proceedings of the event. The publication of abstracts, in the form of a technical report, aims to disseminate the work in progress and completed and record, in a succinct way, the state of the art of the research of the Institute of Computing in 2019.
Abstract: DBpedia is a huge resource available in the Web of Data. It is relevant to update this dataset based on new information appearing in the Wikipedia. However, this operation can provide inconsistencies in the dataset for chapters in different languages. Although existing literature has defined tools to update DBpedia, there is a lack of studies related to understand and detect multi-chapter inconsistencies in its evolution. In this work, we define a set of inconsistency classes related to triple changes in the DBpedia to inform a software tool suited to detect instances of these classes when new data is updated for secondary languages, removed or inserted in the dataset. We conduct an experimental evaluation to assess the types of inconsistencies in several versions of the dataset. We demonstrate that the identified inconsistencies appear in the evolution of DBpedia and propose a solution to correct these inconsistencies. Our evaluation assessing the proposed technique revealed its usefulness. Our results show that the proposed classes of inconsistencies combined with the defined solution can turn the DBpedia more reliable overtime.
Summary: Ontologies and their associated mapping play a central role in various activities related to semantic interoperability in computer systems. However, the constant evolution of ontologies requires updating the existing alignments. Although mapping maintenance techniques have dealt with revisions and removals, adding concepts requires more studies. This research proposes techniques to refine a set of mappings established based on the evolution of ontologies. We investigated ways to suggest correspondences in a new version of the ontology without applying correspondence operations with the entire set of entities of the ontology. The results obtained explore the proximity of the concepts in the alignment process to update the set of mappings. Our experimental evaluation with several versions of alignments between biomedical ontologies shows the effectiveness in considering the context of new concepts.
Summary: Ontologies are artifacts for the representation of knowledge, in which concepts and their interrelations are described within a domain. Ontology networks are characterized by the existence of heterogeneous ontologies in the same conceptual space. A formal or structured ontology is a proposal of determined representation and generally fixed a priori, on the other hand, a “light” ontology is flexible and may be more susceptible to changes in interrelationships over time. The connection between different ontology models is relevant for data integration and the exchange of information between systems. Although techniques exist to automatically relate ontologies by creating correspondences between terms of similar meanings, a model that interconnects heterogeneous ontologies is still an open research topic. This report describes results that allow concepts specified in different model ontologies to be related to each other. We study how formal and well-structured ontologies of Website Semantics can be reconciled with the proposal of less structured “light” ontologies. We obtained a metamodel in eBNF that allows to reconcile structured ontologies described in OWL with “light” ontologies described by RDF triples Fuzzy.
Abstract: This Technical Report presents complementary material related to the article “Identifying and Validating Java Misconceptions Toward a CS1 Concept Inventory”, to be published in the Proceedings of the 24th Annual ACM Conference on Innovation and Technology in Computer Science Education (ITiCSE 2019).
Article Abstract: A misconception is a common misunderstanding that students may have about a specific topic. The identification, documentation, and validation of misconceptions is a long and time-consuming work, usually carried out using iterative cycles of students answering open-ended questionnaires, interviews with instructors and students, exam analysis, and discussion with experts. A comprehensive list of validated misconceptions in some subject can be used to build formal evaluation methods like the Concept Inventory (CI), a multiple-choice questionnaire that is usually performed as pre-post tests in order to assess any change in student understanding. In CS1, validated misconceptions were identified and documented in C and Python programming languages. Although there are studies related to misconceptions in the Java language, these misconceptions lack the formality, comprehensiveness, and robustness of their C and Python counterparts. On this work, we propose a methodology to adapt the validated misconceptions in C and Python to Java. Initially, through the analysis of an initial list of 33 misconceptions in C and 28 in Python, we identified and documented in an antipattern format 31 possible misconceptions in Java. We then developed a final term exam, composed of 7 open-ended questions, in which each question was designed to address some of the misconceptions covered in the course (N = 27). Through the analysis of the exam's answers (N = 69 students), it was possible to validate 22 of the misconceptions (81%). Also, 6 new misconceptions were identified, leading to a total of 28 valid misconceptions in Java.
Abstract: This technical report documents the design and development of efforts that were made in the direction of proposing digital versions of artifacts and techniques used in participatory practices, with the goal of making them accessible to all. Therefore, we aim to allow that any person, regardless of ability, literacy, age, culture or any other special condition, is able to actively contribute and be a part of ideation and participatory practices, used during the design process of new computer systems. In order to accomplish our goal, we adopted the Universal Design (UD) philosophy, so that the artifacts we propose in this work do not require special adaptations for specific groups of users. We also followed the Tangible User Interface (TUI) paradigm, so that our designs can be familiar representations of their original non-digital counterparts, but with an underlying computer system. The artifacts we have redesigned come from Organisational Semiotics (OS) and Participatory Design (PD), and we have done our best to preserve their original essence while making them more inclusive.
Abstract: This technical report presents preliminary results of the museum research scenario developed during the second year of the Socioenactive Systems project. In this report, we first briefly introduce the research scenario, then, we present an overview of our first workshop conducted at the Exploratory Science Museum of Unicamp, in which we exhibited three interactive artworks to be freely explored by participants, who afterwards were invited to build a technological artifact for themselves. Next, we show the preliminary results from the workshop, which include a qualitative analysis of recorded interactions with the three artifacts exhibited and an evaluation of the workshop itself. While the workshop was well evaluated by the participants, the qualitative analysis, in turn, suggests promising socioenactive qualities to be found in the exhibited artifacts, contributing in both theory and practice to the Socioenactive Systems project. We end the report with our concluding remarks and the next steps for the following year.
Summary: Designing interactive computational solutions that effectively help improve human life has been a challenge and a motivation for the Human-Computer Interaction area that requires a deep understanding of the social context and the people who live in that context. Especially in challenging contexts, such as that of a hospital, it is necessary that the design of interactive computational solutions be conducted in a participatory and universal way, both as a process and as a product. In this technical report, we documented three workshops held at the SOBRAPAR hospital, in 2018, with the purpose of investigating and exploring possibilities for socioenative systems that contribute to the quality of life of people in the hospital or that are directly related to it. The workshops were preceded by an ethnographic study, and then were planned and conducted in a participatory and constructive manner in which the results of one workshop informed the planning of the next workshop. In the three workshops, the SobraPet concept, where Pets live in SOBRAPAR, was used as the theme for the workshops with stuffed animals, TV displays and lights to create technological artifacts. The workshops were attended by children and their guardians, in addition to the monitoring of researchers and hospital professionals. As with solutions and dynamics, children's involvement has evolved with each workshop and the results obtained show both the feasibility of the project and inform the planning and execution of future workshops.
Summary: Cross-lingual ontology alignments play a key role for the semantic integration of data described in different languages. The task of automatic cross-lingual ontology matching requires exploring similarities measures. Such measures compute the degree of relatedness between two given terms from ontology's concepts. Although the literature has extensively studied similarity measures for monolingual ontology alignments, the use of similarity measures for the creation of cross-lingual ontology mappings still requires further research. In this work, we define an algorithm for automatic cross-lingual ontology matching based on the analysis of neighbor concepts to improve the effectiveness of the composed similarity approach, a technique to calculate the degree of similarity between concept contents in different languages. Experimental results with OAEI datasets indicate that our novel approach including neighbor concepts for mapping identification has a good effectiveness.
|
Instituto de Computação :: State University of Campinas
Av. Albert Einstein, 1251 - Cidade Universitária Zeferino Vaz • 13083-852 Campinas, SP - Brazil • Phone: [19] 3521-5838 |