TY - JOUR
T1 - Learning to design from humans
T2 - Imitating human designers through deep learning
AU - Raina, Ayush
AU - McComb, Christopher
AU - Cagan, Jonathan
N1 - Funding Information:
This paper has been submitted for publication in the 2019 Design Automation Conference (IDETC). This material is based upon work supported by the Defense Advanced Research Projects Agency through cooperative agreement (Grant No. N66001-17-1-4064; Funder ID: 10.13039/100000185). Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the sponsors.
Publisher Copyright:
Copyright © 2019 by ASME.
PY - 2019/11/1
Y1 - 2019/11/1
N2 - Humans as designers have quite versatile problem-solving strategies. Computer agents on the other hand can access large-scale computational resources to solve certain design problems. Hence, if agents can learn from human behavior, a synergetic human-agent problem-solving team can be created. This paper presents an approach to extract human design strategies and implicit rules, purely from historical human data, and use that for design generation. A two-step framework that learns to imitate human design strategies from observation is proposed and implemented. This framework makes use of deep learning constructs to learn to generate designs without any explicit information about objective and performance metrics. The framework is designed to interact with the problem through a visual interface as humans did when solving the problem. It is trained to imitate a set of human designers by observing their design state sequences without inducing problem-specific modeling bias or extra information about the problem. Furthermore, an end-to-end agent is developed that uses this deep learning framework as its core in conjunction with image processing to map pixel-to-design moves as a mechanism to generate designs. Finally, the designs generated by a computational team of these agents are then compared with actual human data for teams solving a truss design problem. Results demonstrate that these agents are able to create feasible and efficient truss designs without guidance, showing that this methodology allows agents to learn effective design strategies.
AB - Humans as designers have quite versatile problem-solving strategies. Computer agents on the other hand can access large-scale computational resources to solve certain design problems. Hence, if agents can learn from human behavior, a synergetic human-agent problem-solving team can be created. This paper presents an approach to extract human design strategies and implicit rules, purely from historical human data, and use that for design generation. A two-step framework that learns to imitate human design strategies from observation is proposed and implemented. This framework makes use of deep learning constructs to learn to generate designs without any explicit information about objective and performance metrics. The framework is designed to interact with the problem through a visual interface as humans did when solving the problem. It is trained to imitate a set of human designers by observing their design state sequences without inducing problem-specific modeling bias or extra information about the problem. Furthermore, an end-to-end agent is developed that uses this deep learning framework as its core in conjunction with image processing to map pixel-to-design moves as a mechanism to generate designs. Finally, the designs generated by a computational team of these agents are then compared with actual human data for teams solving a truss design problem. Results demonstrate that these agents are able to create feasible and efficient truss designs without guidance, showing that this methodology allows agents to learn effective design strategies.
UR - http://www.scopus.com/inward/record.url?scp=85072533979&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85072533979&partnerID=8YFLogxK
U2 - 10.1115/1.4044256
DO - 10.1115/1.4044256
M3 - Article
AN - SCOPUS:85072533979
VL - 141
JO - Journal of Mechanical Design - Transactions of the ASME
JF - Journal of Mechanical Design - Transactions of the ASME
SN - 1050-0472
IS - 11
M1 - e4044256
ER -