Raciocínio sobre conhecimento visual : um estudo em estratigrafia sedimentar / Reasoning over visual knowledge: a study in sedimentary stratigraphy

AUTOR(ES)
FONTE

IBICT - Instituto Brasileiro de Informação em Ciência e Tecnologia

DATA DE PUBLICAÇÃO

2012

RESUMO

Imagistic domains are those in which problem-solving process begins with the recognition of domain objects trough the visual information captured, supporting more abstract interpretations in subsequent steps. The problem-solving process, in imagistic domains, is performed trough intensive application of visual knowledge, which corresponds to the set of mental models that support the process of reasoning about the information of the spatial arrangement and other visual aspects of the domain entities. The visual knowledge corresponds to a tacit kind of knowledge of practitioners in the field, so it is often used unconsciously, resisting to explicit verbalization. These characteristics of visual knowledge are challenging from the standpoint of Knowledge Engineering. The general aim of this work is to outline an integrated approach to acquisition, modeling, representation and reasoning, to handle visual knowledge, from the standpoint of Knowledge Engineering. The visual interpretation is a common task in imagistic domains, whose resolution demands a reasoning that involves a cognitive process that starts with the direct perception of visual features of objects, and results in abstract understandings of observed scene, such as dynamic behavior of objects in the scene, the meaning of the context of the scene, causes or effects of the state of affairs captured in the scene, etc. Thus, to achieve the overall goal, this work takes as a specific aim to study and model the reasoning process used by experts to solve tasks of visual interpretation, as well as the inferential knowledge structures applied by experts in imagistic domains to accomplish this type of task. The main results of this work are a model of reasoning for solving tasks of visual interpretation; a cognitive grounded model of structure for inferential knowledge representation, called visual chunk; and an approach to visual knowledge acquisition and refinement of ontologies, which explores the reasoning used by the expert as a tool to reveal the lack of important terms in the domain ontology. The study was conducted in the field of sedimentary stratigraphy, focusing on the task of visual interpretation of depositional processes responsible by the generation of sedimentary facies. The models developed in this work was tested in the domain, achieving satisfactory results. In order to apply our approach, a refinement of domain ontology was performed. This process was carried out by applying several techniques of knowledge acquisition in sessions with the expert.

ASSUNTO(S)

engenharia : conhecimento knowledge engineering visual knowledge visualização knowledge models inteligência artificial ontologias cognitive modeling reasoning ontology

ACESSO AO ARTIGO

Documentos Relacionados