Monografias em Ciência da Computação

2018

ABSTRACTS

Departmento de Informática 
Pontifícia Universidade Católica do Rio de Janeiro - PUC-Rio
Rio de Janeiro - Brazil


This file contains a list of the technical reports of the Departmento de Informática, Pontifícia Universidade Católica do Janeiro - PUC-Rio, Brazil, which are published in our series Monografias em Ciência da Computação (ISSN 0103-9741), edited by Prof. Carlos Lucena. Please note that the reports not available for download are available in their print format and can be obtained via the e-mail below.
For any questions, requests or suggestions, please contact:
Rosane Castilho bib-di@inf.puc-rio.br

Last update: 24/MAY/2018

INDEX

[MCC01/18]
 SPYRIDES, G.M.; SANTIAGO, B.; POGGI, M.; LOPES, H. A global integer programming formulation for process discovery. 15 p. Eng. E-mail: poggi@inf.puc-rio.br

Abstract:  Process discovery amounts to determine a process from an event log of a business process. Using the model obtained, one should be able to produce logs from the considered business simulation model. We propose an integer programming formulation that, given a log, determines all places and arcs defining a Petri Net. Formulations from a previous research discover one place at a time. To do so, we extend IPL model in van der Werf et al. 2008 to consider global propreties such as token balance and cohesion among places. Furthermore, the global approach allows more control over the Petri net properties: fitness, simplicity, generalization and precision. We test the resulting methodology on event logs that address most of the pitfalls in processing discovery algorithms. Also, we show the limitations of the method, regarding the Petri net morphology and log scales, and paths for its improvement.

[MCC02/18]
FURTADO, A.L. Notes on arguments: from syllogism to network organized reasoning. 17 p. Eng. E-mail: furtado@inf.puc-rio.br

Abstract: This work is a study on how arguments can be formulated to achieve a better understanding of the factual description provided by a given information system, and to discover alternative sequences of narrative events able to lead to a desired target state of the system. Starting from Aristotle’s syllogism, we proceed by examining problem solving strategies associated with four semiotic relations, which constitute the determinants of information spaces. Next, as a preliminary step to help passing from arguments about facts to arguments concerning events, a prototype is introduced which is able to translate a conceptual specification – involving facts, events and agents – into a relational database implementation equipped with a Log of the executed events. Developed as a logic programming tool with access to Oracle via an ODBC interface, the prototype follows a plan recognition / plan generation paradigm. Within the environment created by the prototype, it becomes possible to extract alternative typical plans, as a process mining task. A network organized reasoning approach is finally proposed as a powerful instrument to analyze the typical plans thus obtained, and trace new plans by traversing event sub-sequences taken from different original alternatives. The same approach is claimed to be applicable to the universe of storytelling, in special to interactively compose new genre-consistent stories by combining existing variants of a folktale.

[MCC03/18]
POCHET, A.; DINIZ, P.H.B.; LOPES, H.; GATTASS, M. Seismic fault detection using convolutional neural networks trained on synthetic post-stacked amplitude maps. 10 p. Eng. E-mail: mgattass@inf.puc-rio.br

Abstract: Fault detection is a crucial step in reservoir characterization. Despite the many tools developed in the past decades, automation of this task remains a challenge. We investigate the application of Convolutional Neural Networks (CNN) to seismic fault detection. CNN is a deep learning method growing in interest in the computer vision community, due to its high performances in a great variety of object detection tasks. One of the constraints of this method is the need to provide a massive number of interpreted data, a requirement particularly difficult to attend in the seismic area. To this end, we built a synthetic dataset with simple fault geometries. The input of our network is the seismic amplitude only; the method does not require computing any seismic attribute. We apply a strategy of patch classification along the images, which requires a simple post process to extract the exact fault location. Our network shows good results on synthetic data and encouraging results when tested on regions of a real section of the Netherland offshore F3 block in the North Sea.

[MCC04/18]
CUNHA, F.J.P.; VIANA, M.L.; SIRQUEIRA, T.F.M.; ROSEMBERG, M.R.; LUCENA, C.J.P. Understanding normative BDI agents behavior. 9 p. Eng. E-mail: lucena@inf.puc-rio.br

Abstract: Testing the autonomy of, and the interaction between, the agents in Multiagent Systems (MAS) is the frontal challenge of traditional software testing approaches. When we study MAS governed by norms – mechanisms created to restrain the behavior of agents – this challenge increases even further. However, agents are autonomous and it is not guaranteed that they will fulfill all norms. Given the fuzzy notion of “test”, especially in the context of MAS, in addition to the difficulties of dealing adequately with normative constraints, the overall understanding of how to handle the creation of tests for normative MAS is still vague. This paper proposes a testing tool to build and run MAS test scenarios and it relies on the use of aspect-oriented techniques to monitor the behavior of autonomous agents. We demonstrated our tool with a simulation of a traffic intersection scenario, based on the Brazilian Transit Code. Our experience shows that the tool can be used to build test scenarios that can achieve high fault detection effectiveness.

[MCC05/18]
CUNHA, F.J.P.; SIRQUEIRA, T.F.M.; VIANA, M.L.; LUCENA, C.J.P. Extending BDI multiagent systems with agent norms. 13 p. Eng. E-mail: lucena@inf.puc-rio.br

Abstract: Open Multiagent Systems (MASs) are societies in which heterogeneous and independently designed entities (agents) work towards similar, or different ends. Software agents are autonomous and the diversity of interests among different members living in the same society is a fact. In order to deal with this autonomy, these open systems use mechanisms of social control (norms) to ensure a desirable social order. This paper considers the following types of norms: (i) obligation — agents must accomplish a specific outcome; (ii) permission — agents may act in a particular way, and (iii) prohibition — agents must not act in a specific way. All of these characteristics mean to encourage the fulfillment of norms through rewards and to discourage norm violation by pointing out the punishments. Once the software agent decides that its priority is the satisfaction of its own desires and goals, each agent must evaluate the effects associated to the fulfillment of one or more norms before choosing which one should be fulfilled. The same applies when agents decide to violate a norm. This paper also introduces a framework for the development of MASs that provide support mechanisms to the agent’s decisionmaking, using norm-based reasoning. The applicability and validation of this approach is demonstrated applying a traffic intersection scenario.

[MCC06/18]
ALMEIDA, A.C.B.; CAMPOS, M.L.M.; SCHWABE, D.; LIFSCHITZ, S. CM-OPL: Configuration Management Ontology Pattern Language Specification. 43 p. Eng. E-mail: sergio@inf.puc-rio.br

Abstract: This document presents the Configuration Management Ontology Pattern Language (CM-OPL). It is the first version of the CM-OPL, represented by using OPLML (Ontology Pattern Language Modeling Language). Therefore, we used a structural model to represent the CM-OPL patterns and structural relationships between them. Also, we present a general process model to provide a general view of the CM-OPL process, and detailed process models expand the process general view.

[MCC07/18]
VELMOVITSKY, P.E.; MILIDIÚ, R.; VIANA, M; LUCENA, C.J.P.; CHAGAS, C.; REHEN, S. iBot: an agent based software framework for creating domain conversational agents. 12 p. Eng. E-mail: lucena@inf.puc-rio.br

Abstract: Conversational Agents (chatbots) are computer programs that interact with users using natural language. Since its inception, the technology has advanced greatly and cloud based platforms from big companies allow developers to create intelligent and efficient chatbots. However, there are not many development approaches to the main modules of a chatbot that are flexible enough to allow the creation of different applications for each domain, while maintaining a robust dialogue control in the application. A promising approach for such a problem is the use of multiagent systems to distribute and perform the tasks performed by the chatbot. This paper introduces a general and flexible framework based on multiagent systems, which will facilitate building chatbots in any domain chosen by the developer, with dialogue control in the application. We conducted a research to outline relevant work and gaps for building a new architecture, which we based our software framework. We show how our approach allows creating applications based on the information state approach to dialogue management in order to increase flexibility and reuse. Our analysis suggests that the tool can be used to build different domains chatbots that allow a robust control of the information in the dialogue. In addition, the use of multiagent systems and the information state approach provided modularity and flexibility to the developed systems.

[MCC08/18]
DAMASCENO, A.L.B.; RIBEIRO, D.S.; COLCHER, S.; BARBOSA, S.D.J. Entrevistas e mapeamento sistemático sobre análise de logs de interação de alunos em ambientes virtuais de aprendizagem. 11 p. Port. E-mail: simone@inf.puc-rio.br

Abstract: Online education has broadened the avenues of research on student's behavior and performance. In this paper, we shed light on how logs of students' behavior in virtual learning environments are analyzed. We conducted interviews with teachers and a systematic mapping of 109 papers on Education Data Mining and Learning Analytics, considering papers published between 2010 and 2018. The main contributions of this paper are on the discussion and analysis of the results that can direct and ground future researches, as the development of tools to evaluate logs of interaction on videos and analyze logs of teachers.

[MCC09/18]
ALMEIDA, A.C.B.; CAMPOS, M.L.M.; BAIÃO, F.; SCHWABE, D.; LIFSCHITZ, S. CM-OPL: configuration management ontology pattern language specification - revised edition. 38 p. Port. E-mail: sergio@inf.puc-rio.br

Abstract: This document presents the Configuration Management Ontology Pattern Language (CM-OPL). It is the second version of the CM-OPL, represented by using OPL-ML (Ontology Pattern Language Modeling Language). Therefore, we used a structural model to represent the CM-OPL patterns and structural relationships between them. Also, we present a general process model to provide a general view of the CM-OPL process, and detailed process models expand the process general view.

[MCC10/18]
CAVALCANTI, R.; PORTUGAL, R.L.Q.; TEIXEIRA, B.; LEITE, J.C.S.P. Em busca dos requisitos para Susana: requisitos para uma humanóide construtora de requisistos. 21 p. Port. E-mail: julio@inf.puc-rio.br

Abstract: In the Silver Jubilee (IEEE Requirements Engineering Conference) track, several authors provided a retrospective of the area, as well as ventured to point out future scenarios for the discipline. Robyn Lutz imagines that by 2042 we'll have humanoids doing requirements engineer’s tasks. Based on this prediction, and using techniques of creativity and elicitation, we are investigating the requirements that this humanoid (called Susana) should have for it to be a requirements generator. This generator would dialogue with people who need to define customized software. In order to identify the requirements of this requirements generator we decided to focus on the kitchen automation domain. Based on this domain, we have done creative work and elicited future demand from possible users of the Humanoid. Those users would demand software through dialogues with Susana. From this demand, we will be infering what requirements would be necessary for building the humanoid. This humanoid, Susana, should generate requirements to be transformed and implemented in kitchens with a high degree of automation. In this report, we describe the initial process used to understand the context of interacting with Susana. Further research will aim which the requirements for building Susana.

[MCC11/18]
 SIRQUEIRA, T.; VIANA, M.; FERNANDES, C.; NASCIMENTO, N.; SARTRE, J.; MIRANDA, P.A.; AUGUSTO, V.; LUCENA, C.J.P. Design de uma plataforma para a gestão de informações médicas distribuídas baseada em sistemas multiagentes e proveniência de dados.  12 p. Port. E-mail: lucena@inf.puc-rio.br

Abstract: This paper discusses the detailed design of a medical information system and proposes an architecture of a distributed information management platform. The system, called Vital Signs, focuses on patients located on Intensive Care Unit (ICU). The Vital Signs is a platform that allows the doctor to monitor in real time the status of patients in the ICU from anywhere, either by a browser or a mobile device. The platform provides a resource for capturing, analysing and monitoring patient data. The concepts underlying the proposed design are based on the features provided by multiagent systems (distributed IA) and the notion of data provenance.