作者: B. Chandrasekaran , Unmesh Kurup , Bonny Banerjee , John R. Josephson , Robert Winkler
DOI: 10.1007/978-3-540-25931-2_16
关键词: Software architecture 、 Artificial intelligence 、 Model-based reasoning 、 Knowledge base 、 Expert system 、 Computer science 、 Component (UML) 、 Diagrammatic reasoning 、 Set (psychology) 、 Matching (graph theory) 、 Diagram 、 Spatial relation
摘要: In problem solving a goal/subgoal is either solved by generating needed information from current information, or further decomposed into additional subgoals. traditional solving, goals, knowledge, and states are all modeled as expressions composed of symbolic predicates, generation rule application based on matching symbols. with diagrams the other hand, an means available, viz., visual perception diagrams. A subgoal opportunistically whichever way successful. Diagrams especially effective because certain types that entailed given explicitly available – emergent objects relations for pickup perception. We add to architecture component representing diagram configuration diagrammatic three basic types, point, curve region; set perceptual routines recognize evaluate generic spatial between objects; action create modify diagram. discuss how domain-specific capabilities can be added top system. The working illustrated scenario.