"Why do people use mental models? First, they are used as inference tools to predict the behavior of a system under novel conditions. They enable us to predict system outcomes from system parameters: We may run our mental model by modifying the system parameters and observing how the behavior of the system changes. Second, mental models can be used to produce explanations and justifications. Such explanations may give us confidence in using e system and enable us to more readily trust the results of the system. Third, mental models can be used as mnemonic devices to facilitate remembering and long-term retention of information. Here, a mental model may provide one with a "cover story" to make the understanding of the system more memorable and easier to recall." (Robbie Nakatsu, "Diagrammatic Reasoning in AI", 1994)
"A hierarchy is a diagram that shows how various components of a system are related, often with a downward direction (or alternatively a left-to-right direction) that moves from more general to more specific. One way to envision a hierarchy is as an inverted tree: We start with a single component (referred to as the root node or topmost node), typically denoted by a square, and then we draw one or more paths leading from it to other nodes. Each of these nodes, in turn, may subdivide into additional subpaths to other nodes. This process may be repeated any number of times to arrive at a multitiered, tree-like structure." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"A mental model is the user's model of a target system; it is a model of a system that exists in a person's head. Through interaction with a complex system, it is a 'naturally evolving model. As a person develops more experience with a system, the model develops and becomes more refined. Hence, at any given point in time, the mental model, as seen through the eyes of the user, is a dynamic, usually incomplete specification of the target system. A conceptual model, on the other hand, is typically the designer's complete specification of a target system. As such, it is intended to be an accurate, consistent, and complete representation of a target system. Ideally, we would want the user's mental model to be the same as the system designer's conceptual model." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Diagrams are information graphics that are made up primarily of geometric shapes, such as rectangles, circles, diamonds, or triangles, that are typically (but not always) interconnected by lines or arrows. One of the major purposes of a diagram is to show how things, people, ideas, activities, etc. interrelate and interconnect. Unlike quantitative charts and graphs, diagrams are used to show interrelationships in a qualitative way." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Models are an important form of knowledge representation because expertise often lies in one's ability to reason about how the objects, or components, of a system are interconnected - whether physically, causally, relationally, or otherwise - in a domain of discourse." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"On the one hand, a conceptual model seeks to faithfully represent the components, the connections, the relations, and the processes that act on the components. On the other hand, a mental model that employs analogical representations is chosen to invite comparisons between two dissimilar domains, never to faithfully and completely represent the target domain." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Prototyping is a method of developing systems rapidly by creating a quick-and-dirty mockup of a system, called a prototype. Once created, the prototype is given to end users so that they can provide their feedback and suggestions for improvement. Based on this feedback, you modify and enhance the prototype. It is an iterative process in that you can get feedback multiple times and enhance the prototype accordingly." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Rule-based expert systems require that you preprogram all the rules that represent the knowledge in the domain. To create a realistic and complete solution. the knowledge engineer must be well versed in the domain and have a clear sense of what the decision procedures are. This understanding must take place beforehand, before the system is created: misunderstandings about the domain can be very costly later on because rule-based expert systems can be extremely difficult to modify and extend into other areas. Even when this is possible. the new rules must be created manually - the expert system does not learn how to tine-tune the rules on its own. Case-based reasoning and neural networks are two Al approaches that are more suitable when you want to create a system that 'learns' how to solve problems on its own." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Semantic networks are used to illustrate how people organize information in their memories. Such representations have been used by cognitive psychologists to understand and theorize how one retrieves and processes information from long-term memory. In AI, semantic networks can also be used as a knowledge representation scheme that programs can use to retrieve information efficiently just like humans do." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Structure is the way that the individual components of a system are interconnected (as given by a system's topology); Behavior refers to what each of these components is supposed to do. From this definition, we may distinguish three levels of system description. First, diagrams are models, graphical in nature, that are used to illustrate structure (e.g., how components are physically interconnected); they do not capture functional behavior of a system. Second, heuristics describe relationships between inputs and outputs, based on the way that experts describe how inputs are transformed into outputs. (Heuristics may be represented as IF-THEN rules). Heuristic knowledge, however, does not attempt to create an explicit representation of system structure. Model-based reasoning is a more complete representation system in the sense that it describes both structure and behavior. From this, three levels of system description can be distinguished, based on whether they describe structure, behavior, or both." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"The development of a mental model, then, can be chronicled, much like the development of a cognitive skill. Three developmental processes8 seem to be at play when a mental model evolves. that mental model becomes more powerful because it works for a wider variety of situations. Second, discrimination means that a mental model is more sensitive to variations in a given situation so that a mental model may add an important new condition where previously it had been overlooked. Third, strengthening means that those aspects of a mental model that have been successfully applied in the past are strengthened and rendered more salient and significant." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"This is always the case in analogical reasoning: Relations between two dissimilar domains never map completely to one another. In fact, it is often the salient similarities between the base and target domains that provoke thought and increase the usefulness of an analogy as a problem-solving tool." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"Venn diagramming, it turns out, is a very effective technique for performing syllogistic reasoning. Its chief advantage (over the Euler graph in particular as we noted earlier) is the ability to incrementally add knowledge to the diagram. While an Euler graph has visual power in terms of representing the relations between sets very intuitively, it is impossible to combine more than one piece of information onto a Euler graph. A Venn diagram, on the other hand, easily lends itself to the representation of partial knowledge and can be manipulated to add successively more knowledge to the diagram. This means that when our knowledge of the relations between sets increases, we simply put in more symbols and shadings into the appropriate compartments of the Venn diagram. Thus we are able to accumulate knowledge in a Venn diagram. This capability turns out to be a powerful feature, one that endows Venn diagrams with a more dynamic quality that is sorely lacking in the Euler system." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
"What advantages do diagrams have over verbal descriptions in promoting system understanding? First, by providing a diagram, massive amounts of information can be presented more efficiently. A diagram can strip down informational complexity to its core - in this sense, it can result in a parsimonious, minimalist description of a system. Second, a diagram can help us see patterns in information and data that may appear disordered otherwise. For example, a diagram can help us see mechanisms of cause and effect or can illustrate sequence and flow in a complex system. Third, a diagram can result in a less ambiguous description than a verbal description because it forces one to come up with a more structured description." (Robbie T Nakatsu, "Diagrammatic Reasoning in AI", 2010)
No comments:
Post a Comment