Marcos A. Castilho

Modal Tableaux with Propagation Rules and Structural Rules

In this paper we generalize the existing tableau methods for modal logics. First of all, while usual modal tableaux are based on trees, our basic structures are rooted directed acyclic graphs (RDAG). This allows natural tableau rules for some modal logics that are difficult to capture in the usual way (such as those having an accessibility relation that is dense or confluent). Second, tableau rules rewrite patterns, which are (schemas of) parts of a RDAG. A particular case of these rules are the single-step rules recently proposed by Massacci. This allows in particular tableau rule presentations for K5, KD5, K45, KD45, and S5 that respect the subformula property. Third, we divide modal tableau rules into propagation rules and structural rules. Structural rules construct new edges and nodes (without adding formulas to nodes), while propagation rules add formulas to nodes. This distinction allows to prove completeness in a modular way.

Petriplan: A New Algorithm for Plan Generation (Preliminary Report)

Recent research shows that Integer Programming techniques are a very promising tool in the AI Planning area. In this paper we define the planning problem as a submarking reachability problem in a Petri net, which is solved by the use of standard IP methods. Using a similar idea to that of the Blackbox algorithm, we show how to translate the plan graph obtained in the first phase of Graphplan into an acyclic Petri net.

On Modelling Virtual Machine Consolidation to Pseudo-Boolean Constraints

Cloud Computing is a new paradigm of distributed computing that offers virtualized resources and services over the Internet. To offer Infrastructure-as-a-Service (IaaS) many Cloud providers uses a large data center which usage ranges 5% to 10% of capacity in average. In order to improve Cloud data center management and resources usage a Virtual Machine (VM) consolidation technique can be applied to increase workloads and save energy. Using VM consolidation, we introduce an artificial intelligence consolidation based in Pseudo-Boolean (PB) Constraints to find a optimal consolidation. To evaluate our PB consolidation approach we used the DInf-UFPR and Google Cluster scenario and the formulas are solved with two state-of-the-art solvers.

Advances in Artificial Intelligence - SBIA 2012

Defeasible Logic Programming (DeLP) is a formalism able to represent incomplete and potentially contradictory information that combines logic programming with defeasible argumentation. In the past few years, this formalism has been applied to real world scenarios with encouraging results. Not withstanding, the outcome one may obtain in this or any other argumentative system is directly related to the decisions (or lack thereof) made during the phase of knowledge representation. In addition, this is exacerbated by the usual lack of a formal methodology able to assist the knowledge engineer during this critical phase. In this article, we propose a formal methodology for knowledge representation in DeLP, that defines a set of guidelines to be used during this phase. Our methodology results in an key tool to improve DeLP’s applicability to concrete domains.

An Investigation on Genetic Algorithms for Generic STRIPS Planning

We investigate the technique of genetic algorithms to solve the class of STRIPS planning problems in Artificial Intelligence. We define a genetic algorithm called AgPlan and we compare it with some recent planners.

Alternância entre competição e colaboração para promover o aprendizado por meio de heurísticas de jogos

Resumo: Em busca de novos metodos que auxiliem no aprendizado de portadores de necessidades especiais, um jogo utilizando a tecnologia de agentes virtuais foi criado, atraves da linguagem brasileira de sinais, LIBRAS, o modelo grafico repassa toda informacao necessaria a manipulacao do jogo didatico, despertando no aluno a autonomia e o interesse pelo conteudo abordado. Abstract: Looking for new methods to help learning, a game using the technology of virtual agent was created, a graphic model using the LIBRAS language to transmit all information that is necessary to the student to manipulate the game, increasing his will of learning and self improving.Resumo: Tem sido ampla e intensamente divulgado e aplicado o uso do computador para fins de aprendizagem em diversos dominios. Partindo dessa perspectiva, deu-se inicio a um trabalho na Casa de Saude Paulo de Tarso, o qual tem como objetivo primordial auxiliar na melhoria da coordenacao motora dos pacientes da referida Instituicao e, por consequencia, aumentar a auto-estima, contribuir com o aprendizado, dentre outros. Abstract: The use of the computer has been wide and intensely divulged and applied in the learning process in diverse dominions. In agreement with this perspective, it was started a work at the Health Institution “Paulo de Tarso”, which has as its major goal to assistant in the improvement of the patients motor coordination in the above-mentioned Institution and, by consequence, to augment the self-esteem, to contribute to apprenticeship, among others.Resumo: A Universidade Federal de Mato Grosso (UFMT) - Campus de Rondonopolis, atraves de alguns estudantes, do 4o ano do curso de Licenciatura em Informatica, em pleno exercicio de seu estagio supervisionado, firmou uma parceria com a EEMOP (Escola Estadual Major Otavio Pitaluga) em maio de 2005 para que juntos realizassem um trabalho, onde os estagiarios estariam auxiliando a utilizacao da Internet durante o processo de ensino-aprendizagem na referida escola. Abstract: Mato Grosso’s Federal University (UFMT) Rondonopolis’s Campus, through some students, of the fourth year of the Graduation Course of Degreee in Informatics, in full exercise of their supervised probation period, established a partnership with EEMOP (Major Otavio Pitaluga State School) in may of 2005 so that together they could accomplish a work, where the trainees would be assisting the use of Internet during the teaching- learning process at the above-mentioned school.Resumo: A educacao a distância (EAD) se apresenta no seculo XXI como um novo espaco de trabalho docente, trazendo mudancas nos papeis tradicionais, requerendo um preparo especial do professor. Para verificar se as IES que oferecem EAD formam seus docentes para ensinar a distância e como fazem este trabalho, foi feita uma pesquisa atraves de questionarios enviados por e- mail para 15 IES de Santa Catarina. O presente artigo discute os resultados encontrados que mostraram que as IES catarinenses nao apenas se preocupam com a producao de materiais didaticos, mas estao criando estruturas de apoio para formar seus professores para ensinar a distância. Abstract: Distance education (DE) is presenting, in the 21st century, a new challenge for teachers, asking for a change in their traditional roles, and requiring them to be specially trained for it. In order to find out if the Higher Education Institutions offering DE are training their teachers to work at a distance and how this preparation has been done, a research was conducted in the state of Santa Catarina, Brazil, using a questionnaire, sent by e-mail to 15 institutions. This article presents and discuss the results, showing that the Higher Education Institutions offering DE are not concerned only about the production of the course materials but are, also, creating support structures to prepare their professionals to teach at a distance.Resumo: Este artigo apresenta resultados iniciais de uma pesquisa visando diagnosticar a situacao da informatica educativa no municipio de Belem(PA). Esse diagnostico foi feito tendo-se como referencia um conjunto de criterios propostos para o planejamento e a avaliacao de experiencias em informatica educativa. Entre os resultados apresentados pode-se destacar o contraste entre a rede publica e a privada e o baixo investimento na formacao docente para essa atividade. Abstract: This paper presents the initial results of a diagnostic of the current state of educative informatics in Belem, Para. To perform such analysis, we used a set of criteria that aimed to evaluate the different educative informatics experiences reported. Among the results it is possible to notice clearly the contrast between the public and private school system and the low investment in the teaching personnel.

Acquiring Expertise in Medical Radiology through Long-Term Interactions

This paper describes how cognitive and computational concepts can be applied to build interface and learner models for long-term tutorial interactions in medical radiology. Key human-to-human tutorial dialogue factors that typically occur in the different stages of skill acquisition are captured through an empirical study. The results of the study are detailed and linked to the design of RUI, an intelligent tutoring system for multiple domains of radiological expertise. A brief discussion and future research directions offer a comparative view of the method and tools.

The Planning Net: Exploring the Petri Net Flow to Improve Planning Solvers

In this paper the classical planning problem is formalized as a Petri Net. We review the Graphplan notions of mutex relation and maintenance actions based on the Petri Net flow. We classify pairs of conflicting actions in terms of four different control structures, which are used to build the Planning Net. The planning problem represented by the net is translated into a SAT instance and then solved using a modern SAT solver. We show the advantages of our method comparing our new planner with Satplan in classical planning domains.

On Modeling Connectedness in Reductions from Graph Problems to Extended Satisfiability

In this paper we present an efficient way to encode connectedness in reductions from graph problems to SAT and MaxSAT. We show and prove linear reductions from Minimum Path and Clique and a quadratic reduction from Steiner Tree, although others NP-complete and NP-hard problems can be reduced with this complexity as well. These reductions use a new class of operators that extends the traditional set of connectors of propositional logic.

Authoring Adaptive Tutoring Systems for Complex Visual Skills

In the world of Intelligent Tutoring Systems research, the connection between learner models and interaction models has been largely ignored. Similarly, previous works have not accounted for the dynamic representation of interface contents based on the underlying pedagogic directives. The paper describes an approach to the authoring of training contents aimed at intermediate-level learners of medical Radiology. It is argued that, in certain dialogues designed for supporting visual diagnosis, it is worthwhile employing domain-general teaching mechanisms and reuse them in various domain-specific situations. An empirical study has been carried out with a corpus of human-to-human tutorial dialogues to identify the component features of expertise in medical Radiology. The results of the study form the basis for implementing interface and learner models that guide long-term tutorial interactions through an intelligent shell called RUI. Conclusions and future research directions are also described briefly.