Bryan Wiltgen
Computer Science Ph.D. Candidate
School of Interactive Computing
College of Computing
Georgia Institute of Technology
Date: September 28, 2018
Time: 2pm – 4:30pm EST
Location: GVU Café, TSRB 2nd Floor
Committee
---
Dr. Ashok K. Goel, School of Interactive Computing, Georgia Institute of Technology (Advisor)
Dr. Nancy Nersessian, School of Interactive Computing, Georgia Institute of Technology
Dr. Mark Riedl, School of interactive Computing, Georgia Institute of Technology
Dr. Spencer Rugaber, School of Computer Science, Georgia Institute of Technology
Dr. Jeannette Yen, School of Biology, Georgia Institute of Technology
Abstract
---
Biologically inspired design is a form of creative analogical design where practitioners take inspiration from nature to develop design solutions, such as using inspiration from owl wings to help reduce the noise made by a train. Although analogical reasoning typically provides no guarantee of correctness and thus evaluation should be an important activity in the domain, there is only sparse existing computational work on supporting evaluation in biologically inspired design. My dissertation develops a computational technique called Design Evaluation through Simulation and Comparison (DESC) to help designers evaluate design concepts in the early, conceptual stage of design.
DESC consists of two methods to evaluate functional models of designs. They are called Simulation and Comparison. The first simulates the model’s processes through qualitative, quantitative, and functional reasoning, and it uses those results to check the model’s claims about process and function. The second analogically compares the model against an alternative model of the same design topic that acts as a proxy for ground truth or an alternative understanding/vision. To support this approach, my dissertation also develops a novel analogical mapping technique called Compositional Mapping that uses problem decomposition and constraint satisfaction.
I evaluated DESC in two ways. I evaluated DESC through computational experimentation that demonstrated Simulation and Comparison’s capability to evaluate functional models of designs, and it provided preliminary evidence that Compositional Mapping is superior to a well-known mapping technique. Additionally, I conducted a pilot study with human participants to test DESC’s usefulness as a support tool. Results provided preliminary evidence that using DESC improves participants’ functional models of design topics.
]]>