Understanding the Relation Between Designer Search Strategies and Designer Learning During Design Space Exploration
Conference Paper
Overview
Research
Identity
Other
View All
Overview
abstract
AbstractDesign Space Exploration (DSE) is a knowledge acquisition technique used from early-stage to detailed engineering design. During DSE, designers systematically generate a range of design alternatives and compare them by their design criteria. DSE allows designers to learn about the design problem, e.g., about design decisions or features that are more common among good designs than among other designs (driving features). To help designers learn the driving features, AI-assisted design tools with advanced analysis techniques have been developed. Yet, there has been scant research attention on understanding designers exploratory and learning behaviors during DSE, which has the potential to contribute to the development of more effective AI-assisted design tools. To address the research gap, we examine designer learning behaviors in real-time using data from a human subject study (N = 24) that studies human-AI collaboration in DSE. We examine design search strategies (i.e., number of generated designs, and scale of design moves made during design generation) and their relations to designer learning (i.e., knowledge about driving features, and ability to use driving features to generate designs). In this work, we show that the designers ability to use driving features to create new designs is associated with their knowledge about the driving features. Also, we observe that designers who generate many designs (mindless creation) achieved lower designer learning than those who generated few designs. We discuss the findings potential implications on the development of future AI-assisted design tools.
