LearnIT: A System That Can Learn and Reuse Design Strategies
Thomas F. Stahovich
Abstract
Thomas F. Stahovich
Abstract
Abstract We describe a system called LearnIT that can observe an iterative solution to a parametric design problem and learn the design strategy employed. The program represents the strategy as a set of rules, which it then uses to automatically generate new design solutions when the design requirements change. Because the rules are learned from the original designer, these new solutions reflect the original designer’s engineering judgment and knowledge of implicit constraints. LearnIT’s approach is based on the observation that often iterative design is actually a form of debugging: each iteration is an attempt to repair a particular flaw in the design. Thus, a program can learn the designer’s strategy by observing what actions are taken in response to each kind of flaw. We have found that the state of the design constraints (satisfied or not satisfied) is a good indicator of what flaws are being addressed at any given time. Because of its ability to capture and reuse the original designer’s understanding of the problem, LearnIT’s primary use is as a design documentation system. However, because it can learn and reuse a design strategy, it can also be considered a design automation tool.
OpenAlex reports 9 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Abstract We describe a system called LearnIT that can observe an iterative solution to a parametric design problem and learn the design strategy employed. The program represents the strategy as a set of rules, which it then uses to automatically generate new design solutions when the design requirements change. Because the rules are learned from the original designer, these new solutions reflect the original designer’s engineering judgment and knowledge of implicit constraints. LearnIT’s approach is based on the observation that often iterative design is actually a form of debugging: each iteration is an attempt to repair a particular flaw in the design. Thus, a program can learn the designer’s strategy by observing what actions are taken in response to each kind of flaw. We have found that the state of the design constraints (satisfied or not satisfied) is a good indicator of what flaws are being addressed at any given time. Because of its ability to capture and reuse the original designer’s understanding of the problem, LearnIT’s primary use is as a design documentation system. However, because it can learn and reuse a design strategy, it can also be considered a design automation tool.
Key concepts: Iterative design, Debugging, Reuse, Computer science, Set (abstract data type), Documentation, Software engineering, High-level design